Decolonizing AI Ethics in Education: A Systematic Review and the Framework for Glocalized AI Ethics in Education (FGAIEE)

Decolonizing AI Ethics in Education: A Systematic Review and the Framework for Glocalized AI Ethics in Education (FGAIEE) | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Decolonizing AI Ethics in Education: A Systematic Review and the Framework for Glocalized AI Ethics in Education (FGAIEE) Dwi Mariyono, Muhammad Yunus, Akmal Nur Alif Hidayatullah This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8417921/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract The rapid integration of artificial intelligence (AI) into educational systems has intensified global efforts to establish ethical guidelines governing its use. However, prevailing AI ethics frameworks in education remain predominantly shaped by Global North epistemologies, universalist moral assumptions, and centralized governance models. As a result, ethical principles frequently fail to translate into contextually legitimate or enforceable practices, particularly in Global South and postcolonial educational settings. This study addresses this gap by critically examining how AI ethics in education is produced, governed, and operationalized across diverse contexts. Using a PRISMA-guided systematic review of 84 peer-reviewed studies published between 2015 and 2025, this study employs an integrated PICo–Thematic synthesis to examine populations, interests, and contexts that are often marginalized in global AI ethics discourse. The analysis reveals three core findings: (1) AI ethics frameworks in education are heavily centralized in Global North institutions, with limited participatory governance and persistent algorithmic bias; (2) pluriversal and localized ethical practices—grounded in Indigenous knowledge systems, communal ethics, and culturally embedded pedagogies—have emerged as viable counter-models; and (3) ethical effectiveness is empirically associated with governance mechanisms that embed ethics into institutional participation, procurement, and accountability structures rather than voluntary principle adoption. Building on these findings, the study advances the Framework for Glocalized AI Ethics in Education (FGAIEE) as its central theoretical contribution. FGAIEE reconceptualizes AI ethics as a multi-scalar governance process that integrates epistemic pluriversality, participatory oversight, and structural enforceability. Rather than proposing another universal ethics model, the framework enables ethical coordination between global AI infrastructures and locally articulated educational values. This study contributes to scholarship on AI ethics, education, and decolonial governance by translating critical theory into an operational framework with global relevance. By repositioning ethics as an issue of epistemic justice and institutional design, FGAIEE offers policymakers, educators, and researchers a pathway to move AI ethics in education from symbolic compliance toward structural redress. AI ethics in education Decolonial governance Glocalized AI ethics Epistemic justice Participatory AI governance Global South perspectives Educational artificial intelligence Systematic review Figures Figure 1 Figure 2 Figure 3 1. Introduction Artificial Intelligence (AI) is rapidly reshaping educational systems worldwide, influencing pedagogical practices, assessment regimes, institutional governance, and policy infrastructures. Framed as a catalyst for personalization, efficiency, and scalability, AI has become embedded in global education reforms—from automated grading and adaptive learning systems to predictive analytics and administrative decision-making. Yet, as AI technologies proliferate across educational contexts, the ethical frameworks governing their design and deployment remain overwhelmingly shaped by Euro-American epistemologies and Global North policy imaginaries. This imbalance raises a critical question: ethical for whom, by whom, and under what epistemic authority ? Contemporary AI ethics in education is dominated by universalist frameworks produced by international organizations and Global North institutions, such as the European Union’s Ethics Guidelines for Trustworthy AI, UNESCO’s AI and Education recommendations, and the OECD AI Principles (Floridi et al., 2018 ; Jobin et al., 2019 ; OECD, 2023 ; UNESCO, 2021 ). While these initiatives have significantly advanced global awareness of fairness, transparency, and accountability, they are typically formulated within contexts of robust digital infrastructure, Anglophone datasets, and liberal-individualist moral traditions. When exported as “global standards,” these frameworks often fail to account for the cultural, linguistic, political, and infrastructural realities of the Global South, thereby reproducing what critical scholars describe as digital colonialism (Arora et al., 2023 ; Couldry & Mejias, 2019 ; Kwet, 2019 ). Existing frameworks such as the community-wide ethics model for AIED (du Boulay, 2023 ; Holmes et al., 2022 ) and studies on AI ethics education (Wiese et al., 2025 ) provide foundational principles for fairness and accountability in educational AI, but they remain largely descriptive and insufficiently connected to governance structures that account for epistemic asymmetries across diverse sociocultural contexts. In educational settings, the consequences of this imbalance are increasingly evident. AI-driven assessment systems systematically underperform for non-native English speakers; algorithmic decision-making privileges dominant cultural norms embedded in training data; and governance mechanisms rarely provide meaningful participatory space for educators, learners, or policymakers from marginalized regions (Baker & Hawn, 2022 ; Heeks, 2022 ; Zajko, 2021 ). Ethical safeguards designed to promote equity thus risk entrenching inequality—transforming AI ethics from a protective instrument into a vehicle of epistemic exclusion. This phenomenon reveals a deeper problem: AI ethics in education is not merely a technical or normative domain, but a site of epistemic power and geopolitical struggle . Despite a growing body of scholarship on AI ethics and digital colonialism, three critical gaps persist. First , existing AI ethics literature overwhelmingly treats ethics as a universal, value-neutral checklist, insufficiently interrogating whose epistemologies define ethical legitimacy and how power asymmetries shape “global” standards ((Floridi et al., 2018 ; Jobin et al., 2019 ). Second , systematic reviews in this field tend to focus either on technical fairness metrics (Mehrabi et al., 2022 ) or on high-level mappings of ethical principles, without offering integrative methodological approaches capable of translating critical insights into actionable governance mechanisms. Third , global AI ethics frameworks in education remain largely principle-based, resulting in symbolic compliance rather than enforceable structural redress—particularly in contexts marked by infrastructural constraints and epistemic marginalization (Sahebi & Formosa, 2024 ). To address these gaps, this study reconceptualizes AI ethics in education through the lenses of epistemic pluriversality and glocalized governance . Drawing on decolonial theory, pluriversality recognizes multiple knowledge traditions—such as Ubuntu ethics, Indigenous data sovereignty, Islamic epistemologies, Confucian relational ethics, and gotong royong governance—not as cultural add-ons, but as coequal sources of ethical authority (Mignolo, 2020 ; Nurhafida et al., 2020 ; Zembylas, 2023 ). Glocalized governance, in turn, rejects the unidirectional export of global norms and instead embeds ethical principles within locally determined practices, ensuring that AI systems are co-designed, audited, and governed by the communities they affect (Abdalla et al., 2021 ; Olojede, 2023 ; UNESCO, 2021 ). Methodologically, this study conducts a PRISMA-guided systematic review of 84 peer-reviewed studies published between 2015 and 2025, employing an original PICo–Thematic–SWOT synthesis . This multi-layered approach integrates: (1) the PICo framework to foreground populations, interests, and contexts that are often marginalized; (2) thematic analysis to surface epistemic patterns and ethical tensions; and (3) SWOT analysis as an epistemic diagnostic tool that translates critical findings into policy-relevant leverage points. Unlike conventional reviews, this design moves beyond description to generate strategic pathways for ethical governance in education. The study is guided by three research questions: RQ1 (Diagnostic – Coloniality as Mechanism) : How do dominant AI ethics frameworks in education operationalize assumptions of universal ethical transferability, and through what sociotechnical mechanisms—such as datasets, procurement regimes, benchmarking practices, and accreditation systems—do they reproduce geopolitical and epistemic asymmetries between the Global North and the Global South? RQ2 (Analytic – Epistemic Disobedience & Pluriversality) : What forms of epistemic disobedience and pluriversal ethical practices have emerged within educational AI systems across diverse contexts, and how do these practices reconfigure participation, accountability, and ethical legitimacy beyond universalist frameworks? RQ3 (Constructive – From Principles to Governance) How can AI ethics in education be reconceptualized from principle-based or symbolic compliance toward structurally inclusive and glocalized governance models that institutionalize epistemic pluriversality across policy, design, and implementation? In response to these questions, the study advances its central contribution: the Framework for Glocalized AI Ethics in Education (FGAIEE) . FGAIEE reconceptualizes ethical AI governance as a multi-scalar, participatory process that bridges global infrastructures with local epistemologies through enforceable mechanisms such as bias-linked procurement, participatory audits, South-led open-source toolkits, and culturally responsive teacher training. In doing so, this article positions AI ethics in education not as a universal moral template, but as a political project of epistemic justice and shared authorship—one that will ultimately shape whether AI entrenches or transforms global educational futures. Rather than presenting a separate theoretical framework section, this study integrates critical and decolonial perspectives into the introduction and develops them through empirical synthesis and theoretical construction in the Discussion. 2. Method 2.1 Research Design This study adopts a PRISMA-guided systematic literature review to critically examine AI ethics in education through a decolonial and governance-oriented lens (Page et al., 2021 ). Unlike conventional AI ethics reviews that primarily synthesize ethical principles or normative positions (Floridi et al., 2018 ; Jobin et al., 2019 ), this review is designed to interrogate how ethical authority, epistemic power, and governance logics are constructed, circulated, and contested across global educational contexts. To address this epistemically complex problem, the study employs a layered methodological architecture integrating PICo framing, thematic analysis, and SWOT synthesis. This design reflects calls within critical AI studies to move beyond descriptive aggregation toward explanatory and governance-relevant synthesis (Heeks, 2022 ). The methodological sequence allows the analysis to progress from contextual specification (PICo), to interpretive pattern detection (thematic analysis), and finally to policy-operable synthesis (SWOT). 2.2 Review Protocol and Search Strategy The review followed the PRISMA 2020 statement to ensure transparency, rigor, and replicability in study identification, screening, and inclusion (Page et al., 2021 ). Searches were conducted across Scopus, Web of Science, ERIC, and Google Scholar, which collectively capture dominant, interdisciplinary, and education-focused scholarship. Search strings combined terms related to artificial intelligence , algorithmic systems , ethics , governance , education , equity , Global South , and decolonial perspectives , using Boolean operators to balance breadth and specificity. In line with critiques of epistemic exclusion in AI research (Couldry & Mejias, 2019 ; Mohamed et al., 2020 ), the search strategy intentionally avoided limiting results to formal AI ethics frameworks, enabling the inclusion of studies addressing Indigenous data sovereignty, pluriversal ethics, and localized governance practices. 2.3 Inclusion and Exclusion Criteria Studies were included if they: Examined AI or algorithmic systems in educational contexts (K–12, higher education, lifelong learning, or digital learning platforms); Engaged explicitly with ethical, governance, justice, or power-related issues; Were peer-reviewed journal articles published between 2015 and 2025, reflecting the maturation of AI ethics discourse; and Provided empirical, theoretical, or policy-relevant insights. Studies were excluded if they: Focused solely on technical optimization without ethical or governance analysis (Mehrabi et al., 2022 ); Addressed AI ethics outside education; Were editorial commentaries without analytical grounding; or Lacked sufficient methodological transparency. Following screening and eligibility assessment, 84 studies were retained for final analysis ( Fig. 1 ) . 2.4 PICo Framework for Analytical Structuring The PICo framework —Population, Interest, and Context—was employed to structure the review and maintain analytical coherence across heterogeneous studies (Stern et al., 2014 ). Population (P) : Educational actors and systems, including learners, educators, institutions, and policymakers across Global North and Global South contexts. Interest (I) : Ethical governance of AI in education, with emphasis on epistemic authority, algorithmic bias, participation, and accountability. Context (Co) : Sociotechnical, cultural, and geopolitical settings in which AI ethics frameworks are developed, implemented, and legitimized. PICo was used not merely as a categorization tool, but as an interpretive scaffold that foregrounds context as analytically central—responding to critiques that ethics frameworks often treat educational settings as neutral implementation spaces (Baker & Hawn, 2022 ; Zajko, 2021 ; Zembylas, 2023 ). 2.5 Data Extraction and Thematic Analysis Data extraction captured publication characteristics, ethical frameworks referenced, educational contexts, governance mechanisms, and reported ethical tensions or outcomes. A hybrid inductive–deductive thematic analysis was conducted following the reflexive approach outlined by Braun and Clarke ( 2006 , 2021 ). Deductive coding was informed by critical AI ethics and decolonial scholarship, including concepts such as universality , epistemic justice , digital colonialism , and participatory governance (Birhane, 2021 ; Kwet, 2019 ; Mignolo, 2020 ). Inductive coding allowed unanticipated patterns—particularly counterintuitive and tension findings—to emerge from the data. This approach ensured theoretical sensitivity without constraining analytical openness. 2.6 Inter-Rater Reliability and Analytical Rigor To enhance analytical rigor and credibility, coding was conducted independently by multiple reviewers. Inter-rater reliability was assessed using Cohen’s kappa, (Appendix Table 3B) yielding agreement levels categorized as excellent according to established benchmarks (McHugh, 2012 ). Discrepancies were resolved through iterative dialogue, prioritizing interpretive alignment over mechanical consensus. Such procedures are particularly important in reviews engaging with normative and epistemic questions, where analytical transparency is essential to scholarly trustworthiness (O’cathain et al., 2008 ). 2.7 SWOT Synthesis for Governance-Oriented Interpretation While thematic analysis captured epistemic patterns and power relations, a further analytical step was required to translate critical insights into governance-relevant implications . To this end, a SWOT synthesis was employed at the interpretive stage. Consistent with prior critical policy reviews (Heeks, 2022 ; Helms & Nixon, 2010 ), SWOT was repurposed not as a managerial tool but as a structuring device to connect epistemic critique with policy operability. Strengths and weaknesses reflected internal characteristics of dominant AI ethics frameworks, while opportunities and threats captured external geopolitical, institutional, and cultural dynamics shaping educational AI governance. 2.8 Methodological Justification and Reflexivity The integration of PICo framing, PRISMA-guided screening, thematic analysis, and SWOT synthesis is not methodological excess, but a deliberate response to a recognized limitation in existing AI ethics reviews. Much of the literature remains confined to principle-level comparison or descriptive synthesis, offering limited insight into how ethical authority is operationalized within real governance infrastructures (Birhane, 2021 ; Floridi et al., 2018 ; Jobin et al., 2019 ). Methodological simplification would risk reproducing the very universalist assumptions this review seeks to interrogate. By contrast, methodological plurality enables the study to bridge epistemic critique with policy-operable governance design , addressing a gap widely acknowledged yet rarely resolved in AI ethics and education scholarship (Couldry & Mejias, 2019 ; Heeks, 2022 ; Mohamed et al., 2020 ). 3. Results This section reports the empirical findings of the systematic review of 84 peer-reviewed studies (2015–2025). Results are organized according to the three research questions (RQ1–RQ3) and present observed patterns, distributions, and documented practices derived from the data. Interpretive, theoretical, and normative analysis is reserved exclusively for the Discussion section. Two integrative results artifacts anchor this section: Table 2 , which synthesizes dominant empirical themes, representative studies, and observed policy implications; and Table 3 , which maps distributional patterns of AI ethics in education through a decolonial analytical lens. Together, these tables serve distinct but complementary empirical functions: Table 2 synthesizes recurring thematic findings across the literature, while Table 3 maps how these themes are unevenly distributed across geopolitical contexts, datasets, and governance arrangements. 3.1. RQ1. How do current global AI ethics frameworks in education reproduce or challenge colonial epistemologies and structural asymmetries? The review reveals a pronounced geopolitical concentration in the production and circulation of AI ethics frameworks applied to education . Across the dataset, 78% of identified frameworks and policy instruments originated from Global North institutions , primarily Europe and North America (Khan et al., 2022 ; Pasupuleti, 2024 ; Upadhyay et al., 2023 ). These frameworks were most often implemented in Global South contexts through policy transfer, donor requirements, or procurement conditions, with limited evidence of local co-authorship. Empirical studies further document systematic algorithmic performance disparities . AI-driven assessment and language-processing systems exhibited accuracy gaps of approximately 10–15% for non-native English speakers , compared with native speakers, across multiple educational settings (Baker & Hawn, 2022 ; Zajko, 2021 ). These disparities were consistently associated with English-dominant training datasets and culturally narrow pedagogical assumptions. Governance and participation patterns mirrored these asymmetries. Fewer than one in five studies reported mechanisms enabling educators, learners, or policymakers from marginalized regions to participate meaningfully in ethical decision-making processes related to AI adoption (Arora et al., 2023 ; Sahebi & Formosa, 2024 ). Where participation was noted, it was predominantly advisory rather than deliberative or decision-making in nature. These empirical patterns are synthesized thematically in Table 2 , which consolidates recurrent findings related to Global North dominance, algorithmic exclusion, and participatory gaps across studies. Table 3 , by contrast, maps the distributional structure of these findings, demonstrating how authorship, datasets, and governance authority are disproportionately concentrated in Global North contexts. 3.2. RQ2. What pluriversal counter-strategies and localized practices have emerged to foster epistemic inclusion and ethical legitimacy? Alongside dominant global frameworks, the review identified a growing set of context-specific ethical practices emerging from Global South, Indigenous, and localized educational settings. These practices were characterized by participatory governance, cultural embeddedness, and local accountability mechanisms. Several studies documented Ubuntu-informed co-design models in African educational contexts. In Ghana, participatory ethical deliberation grounded in communal decision-making was associated with a reported 40% increase in teacher engagement and trust toward AI-supported educational tools (Ammah et al., 2024 ). Comparable participatory models rooted in gotong royong principles were reported in Southeast Asian contexts, emphasizing collective responsibility for AI oversight and evaluation (Madlberger, 2017 ). Another recurring pattern involved Indigenous data sovereignty and custodial governance arrangements . Studies reported models in which data ownership, consent, and usage were negotiated collectively, positioning educational data as a shared resource rather than an extractable commodity (Zembylas, 2023 ). These approaches were frequently linked to heightened transparency and community trust. Additionally, the review identified South-led open-source and frugal AI initiatives as operational counter-strategies to infrastructural dependency. These initiatives prioritized multilingual design, low-bandwidth functionality, and local capacity-building, enabling ethical adaptation in resource-constrained environments (Ng et al., 2021b , 2021a ). These pluriversal practices are consolidated empirically in Table 2 as recurring thematic patterns with documented outcomes. Table 3 situates these practices within the broader global distribution of AI ethics research , illustrating their relative marginality compared to dominant Global North frameworks. 3.3. RQ3. What empirical pathways indicate movement from symbolic ethical compliance toward structural redress in AI governance for education? The reviewed studies consistently distinguished between symbolic adoption of ethical principles and governance mechanisms capable of producing observable structural effects . While global AI ethics guidelines—most notably UNESCO (UNESCO, 2021 )—were frequently cited, fewer than 25% of studies documented enforceable mechanisms linking ethical commitments to institutional accountability (Sahebi & Formosa, 2024 ). Empirical synthesis identified three governance pathways associated with more substantive outcomes: Bias-linked procurement and auditing, where ethical compliance was tied to algorithmic performance evaluations and contractual obligations (Nazer et al., 2023 ; Upadhyay et al., 2023 ). Institutionalized participatory governance, including educator-led review boards and community audit mechanisms, which were associated with improved trust and contextual relevance (Olojede, 2023 ). Capacity-building through AI literacy and open infrastructures, enabling sustained local adaptation and oversight (Madlberger, 2017 ; Ng et al., 2021b , 2021a ). These pathways are not presented as normative prescriptions but as empirically observed mechanisms across the reviewed studies. Table 3 captures how such mechanisms remain unevenly distributed across regions, reinforcing the contrast between symbolic compliance and structurally embedded ethics. 3.4. Integrative Empirical Synthesis Table 2 functions as the primary thematic synthesis of empirical findings, identifying what ethical issues and governance practices recur across AI ethics in education. Table 3 complements this synthesis by providing a distributional analysis , mapping where these themes are produced, implemented, and governed, and by whom. Together, the tables distinguish between the content of ethical concerns and the structural conditions under which they are institutionalized. Table 2 Thematic synthesis of AI Ethics in Education: From Empirical Findings to Policy Implications Theme Representative Studies Theoretical Link Policy Implication Dominance of Global North frameworks in AI ethics Khan et al. ( 2022 ); Upadhyay et al. ( 2023 ) Coloniality of knowledge : Euro-American standards universalized as “global” Require glocal adaptation of guidelines; reserve policy seats for Global South actors Algorithmic exclusion of non-native speakers Zajko ( 2021 ); Baker & Hawn ( 2022 ) Colonial infrastructures : monolingual datasets reproduce systemic bias Mandate multilingual datasets and local bias audits in procurement Lack of participatory governance UNESCO ( 2021 ); Olojede ( 2023 ) Epistemic disobedience : absence of local voices shows systemic silencing Institutionalize community co-design and participatory policy mechanisms Ubuntu-informed and Indigenous approaches Ammah et al. ( 2024 ); Zembylas ( 2023 ) Pluriversality : multiple knowledge systems as coequal ethical sources Scale up South-led, culturally grounded AI ethics toolkits Vendor lock-in and dependency Arora et al. ( 2023 ); Sahebi & Formosa ( 2024 ) Coloniality of power : economic dependency embedded in digital infrastructures Enforce open-source alternatives and transparent procurement Culturally responsive pedagogy in AI literacy Madlberger (2017); Vetter (2024) Epistemic justice : validating local pedagogical traditions Integrate AI ethics into teacher training and curricula in context-sensitive ways Table 3 captures how these governance mechanisms remain unevenly distributed across regions, highlighting structural disparities rather than thematic repetition. Table 3 Distributional Mapping of AI Ethics in Education Studies through a Decolonial Lens Theme Representative Studies Theoretical Link (Decolonial Lens) Policy Implication Geopolitical dominance of Global North Khan et al. ( 2022 ) – majority frameworks Euro-American; Upadhyay, Pasupuleti ( 2024 ) – weak Global South representation Coloniality of knowledge : Euro-American epistemologies universalized as “global” Develop South-led frameworks; allocate research funding to underrepresented regions Algorithmic exclusion & linguistic bias Zajko ( 2021 ); Baker & Hawn ( 2022 ) – ~15% accuracy gap for non-native English speakers Coloniality of language : English-centric datasets systematically marginalize others Require multilingual datasets; enforce context-aware validation of algorithms Tokenistic inclusion without participation UNESCO ( 2021 ) – guidelines often adopted symbolically; Sahebi & Formosa ( 2024 ) – superficial compliance Epistemic disobedience absent → participation remains formal, not substantive Institutionalize stakeholder quotas; mandate community co-design in EdTech projects Emerging pluriversal counter-models Ammah et al. ( 2024 ) – Ubuntu AI ethics; Madlberger (2017)– localized pedagogies Pluriversality : diverse epistemologies as legitimate ethical foundations Scale up co-created curricula; fund pluriversal AI literacy and governance toolkits The complete table is presented in the Appendix ( Table 3 A ). This table presents a distributional analysis of empirical findings across the reviewed studies, mapping where AI ethics frameworks in education are produced, whose epistemic perspectives dominate, and how governance authority is allocated across geopolitical contexts. Unlike the thematic synthesis in Table 2 , which consolidates recurrent ethical issues and practices, Table 3 highlights structural asymmetries in authorship, datasets, linguistic orientation, and decision-making power. The table provides empirical evidence of how ethical principles and governance mechanisms are unevenly institutionalized, revealing persistent concentrations of authority in Global North contexts alongside the relative marginality of pluriversal and locally grounded approaches. Results Summary In summary, the Results indicate that: AI ethics in education is predominantly shaped by Global North–produced frameworks with limited contextual adaptation. Algorithmic bias and governance exclusion are empirically documented and systematically patterned. Pluriversal and localized practices constitute observable ethical alternatives rather than isolated exceptions. Structural redress is empirically associated with governance mechanisms that embed ethics into participation, procurement, and capacity-building processes. Interpretation of these findings, including their theoretical implications and the development of the Framework for FGAIEE , is presented in the following Discussion section. 4. Discussion Recent scholarship on AI ethics in education has highlighted the limitations of principle-based and universalist frameworks. Although influential models have established shared vocabularies of fairness, transparency, and accountability (du Boulay, 2023 ; Holmes et al., 2022 ), empirical evidence shows that these principles often remain weakly connected to institutional governance and pedagogical realities, particularly beyond the Global North. Consistent with the findings of this review, ethical commitments alone rarely translate into enforceable or contextually legitimate practices in educational AI systems. Taken together, these findings reposition AI ethics in education as a governance problem rooted in epistemic power and institutional design, rather than a deficit of ethical principles. While prior work in AI and education has addressed ethical principles and pedagogical competence (Holmes et al., 2022 ; Wiese et al., 2025 ), governance-focused reviews reveal that ethical guidelines often lack enforceable mechanisms (Batool et al., 2025 ). These observations align with our findings and underscore the need for frameworks that bridge global principles with local legitimacy — a gap addressed by the FGAIEE . 4.1. Reframing AI Ethics in Education as an Epistemic Power Structure The findings reported in RQ1 confirm that AI ethics in education operates not merely as a normative or technical domain, but as an epistemic power structure . The dominance of Global North–produced ethical frameworks reflects what Quijano ( 2000 ) conceptualized as the coloniality of power , wherein authority over knowledge production persists beyond formal colonialism (Newman, 2024 ). When ethical legitimacy is defined externally—through Euro-American moral philosophies, Anglophone datasets, and universalist governance templates—AI ethics risks reproducing epistemic hierarchies under the guise of neutrality. This observation aligns with Santos’ (2023) critique of epistemicide , in which alternative knowledge systems are rendered invisible by claims of universality. The Results demonstrate that global AI ethics frameworks in education often function as regulatory abstractions, disconnected from lived pedagogical realities. As evidenced by the lack of participatory governance and persistent algorithmic bias, ethics becomes procedural rather than transformative. Table 4 deepens this interpretation by demonstrating how principle-based ethics frameworks translate into symbolic compliance precisely because they are decoupled from decision-making authority and enforcement mechanisms. The table illustrates that ethical commitments remain largely aspirational unless embedded in enforceable institutional mechanisms. This finding resonates with critiques of “ethics washing” in AI governance (Bietti, 2020 ; Papyshev & Chan, 2025 ), extending them specifically into educational contexts. Table 4 Practical Leverage Points for Ethical AI in Education Leverage Point Illustrative Examples / Literature Integration with FGAIEE (1) 50% representation of local communities in AI design Gotong royong deliberation; Ubuntu ethics; vanua governance Ensures epistemic inclusion in governance processes (2) Mandatory bias audits & diverse data engagement Bias reduction initiatives (Nazer et al., 2023 ) Embeds accountability and transparency into procurement (3) Public–private partnerships for low-cost AI in LMICs Infrastructure pilots in resource-constrained schools Links equity to resource redistribution (4) Integration of AI literacy into curricula & teacher training Four-pillar approach (Ng et al., 2021a , b ) Promotes ethical awareness as a systemic outcome (5) Local culture-based participatory design Digital storytelling (Davy Tsz Kit et al., 2022 ); ethical co-design (Vetter et al., 2024 ) Embeds cultural values as ethical foundations (6) South-led AI ethics research funding Increase in Global South publications and grants Strengthens epistemic sovereignty in knowledge production (7) Glocalized governance model UNESCO’s ROAM-X pillars adapted locally (Papakostas, 2025 ) Operationalizes global–local co-ownership of standards (8) Inclusion- and disability-friendly design (frugal AI) SMS-based AI, local-language platforms, inclusive UI Ensures accessibility as a marker of ethical AI From Universalism to Pluriversality The empirical patterns identified in RQ2 provide strong support for pluriversality as a necessary theoretical shift in AI ethics. Drawing on Mignolo ( 2020 ), pluriversality rejects the assumption that ethical universals can be derived from a single epistemic center. Instead, it affirms the coexistence of multiple ethical rationalities. Where participatory governance concerns who decides , epistemic pluriversality addresses whose knowledge defines what counts as ethical in the first place . These findings align with emerging Global South–centered scholarship that foregrounds communal and relational ethics in AI governance. Studies in African higher education contexts emphasize Ubuntu as a normative foundation for AI ethics, positioning ethical responsibility as collective stewardship rather than individual compliance (Corrigan et al., 2023 ; Yilma, 2025 ). Such perspectives empirically reinforce the argument that ethical legitimacy in education is culturally embedded and epistemically plural. Practices grounded in Ubuntu ethics, gotong royong , and Indigenous data sovereignty exemplify what Zembylas ( 2023 ) describes as relational ethics —ethical reasoning embedded in social responsibility, reciprocity, and collective care. These findings challenge dominant liberal-individualist ethics frameworks that prioritize autonomy and consent detached from community context. Figure 2 functions as an analytical pivot, translating the empirical patterns identified in RQ2 into a coherent model of how local ethical traditions interact with global AI infrastructures. Rather than positioning local ethics as subordinate or supplementary, the figure conceptualizes them as co-constitutive layers of ethical governance. This reframing moves beyond cultural “adaptation” toward epistemic parity. The dialog between Results and theory here confirms that pluriversal ethics are not merely philosophical ideals but empirically observable governance practices . As such, they provide a credible foundation for rethinking AI ethics in education at scale. This figure synthesizes empirical patterns identified across RQ1–RQ3, illustrating the transition from principle-based and symbolic AI ethics toward structurally embedded governance mechanisms. The model highlights how participatory governance, epistemic inclusion, and institutional accountability function as empirical pathways enabling ethical principles to translate into legitimate and enforceable practices in educational AI systems. Operationalizing Ethics: From Principles to Governance Mechanisms One of the most significant contributions of this study lies in clarifying the distinction between ethical principles and ethical mechanisms . The Results reveal that while ethical principles are widely cited, they rarely translate into enforceable institutional action. This gap echoes Heeks’ ( 2022 ) argument that digital development initiatives often fail due to an “implementation void” between intention and practice. Table 5 plays a critical role in bridging this gap by identifying empirically grounded governance instruments—such as bias-linked procurement, participatory audits, and educator-led oversight—that operationalize ethical commitments. Table 5 does not introduce new empirical findings but synthesizes observed disparities into a comparative governance heuristic that clarifies pathways for structural enforceability. These mechanisms resonate with Ostrom’s ( 1990 , 2010 ) and Carlisle & Gruby ( 2019 ) theory of polycentric governance , which emphasizes distributed authority and collective rule-making over centralized control. By situating ethics within procurement contracts, audit procedures, and institutional accountability structures, these mechanisms shift AI ethics from moral persuasion to structural enforceability . This distinction is essential for education systems, where AI tools directly influence assessment, access, and learner trajectories. Table 5 Heatmap: Regional Disparities in AI Ethics Implementation Region Infrastructure Readiness Policy Strength Cultural Adaptability Stakeholder Inclusion Sub-Saharan Africa Low (●○○) Medium (●●○) High (●●●) Low (●○○) South Asia Medium (●●○) Low (●○○) Medium (●●○) Medium (●●○) EU/North America High (●●●) High (●●●) Low (●○○) High (●●●) Latin America Medium (●●○) Medium (●●○) High (●●●) Low (●○○) Explanation : • ●●● = High/Strong, ●●○ = Medium, ●○○ = Low Source: Synthesis of Yadav et al. ( 2025 ), Wakunuma & Eke ( 2024 ), and Shams et al. ( 2023 ). Insight : Global North-South Pattern: Strong infrastructure and policy readiness in the North, better cultural adaptation in the South. Critical Gap: Low stakeholder inclusion in the Global South despite high cultural adaptation. Similar critiques have been articulated in recent work on decolonising educational technology, which argues that ethical AI in education requires structural reconfiguration of governance, not merely ethical adaptation of existing technologies (Koole et al., 2024 ). 4.2. The FGAIEE Framework: Toward Glocalized and Enforceable AI Ethics in Education The empirical findings of this review indicate that prevailing AI ethics frameworks in education struggle to translate universal ethical principles into contextually legitimate and enforceable governance practices. As demonstrated in the Results (RQ1–RQ3), ethical failures are not incidental but structurally produced through epistemic centralization, limited participation, and weak accountability mechanisms. In response, this study advances the FGAIEE as an integrative theoretical synthesis that bridges empirical evidence with decolonial and governance theory. FGAIEE is grounded in the proposition that ethical legitimacy in AI-mediated education emerges through negotiated alignment between global infrastructures and local epistemologies , rather than through the uncritical adoption of universal norms. This positioning extends decolonial critiques of digital colonialism (Quijano, 2000 ; Santos, 2014; Mignolo, 2020 ) by translating them into an operational framework applicable to educational governance. Epistemic Pluriversality as the Ethical Foundation The first pillar of FGAIEE is epistemic pluriversality , which recognizes multiple knowledge systems as coequal sources of ethical authority. The Results demonstrate that Global North–centric ethics frameworks frequently marginalize local pedagogical values, linguistic norms, and culturally embedded assessment practices. FGAIEE responds by repositioning local ethical traditions—such as Ubuntu ethics, gotong royong , Indigenous data sovereignty, and religious epistemologies—not as contextual adaptations, but as foundational inputs into ethical governance. As illustrated in Fig. 2 , epistemic pluriversality operates as the base layer of the framework, shaping how ethical questions are defined, deliberated, and resolved. This approach aligns with relational ethics perspectives (Zembylas, 2023 ) and directly addresses the epistemic exclusion observed across the reviewed studies. Rather than seeking ethical uniformity, FGAIEE enables ethical coordination without epistemic domination . Participatory and Polycentric Governance The second pillar operationalizes ethics through participatory and polycentric governance structures . Consistent with the empirical patterns identified in RQ1 and RQ2, the absence of meaningful stakeholder participation correlates with ethical fragility and low institutional trust. Drawing on Ostrom’s ( 1990 ) theory of polycentric governance, FGAIEE distributes ethical authority across multiple actors and levels, including educators, learners, communities, institutions, and policymakers. The centrality of educator participation identified in this study is consistent with recent empirical research showing that educators perceive AI ethics as a complex, context-dependent challenge requiring institutional support rather than abstract guidance (Kamali, 2024 ). This reinforces the framework’s emphasis on participatory and polycentric governance as a prerequisite for ethical sustainability in educational AI systems. This governance logic is reflected in Table 4 , which synthesizes empirically documented mechanisms such as educator-led ethics boards, community review processes, and participatory audits. These mechanisms demonstrate that ethical legitimacy increases when those affected by AI systems are directly involved in their oversight. Within FGAIEE, participation is not consultative but decision-oriented , ensuring that ethical governance remains context-sensitive and contestable. Structural Enforceability of Ethical Commitments The third pillar addresses the most persistent weakness of existing AI ethics frameworks: the lack of enforceability. The Results reveal that ethical principles frequently remain symbolic unless embedded within institutional mechanisms. FGAIEE therefore conceptualizes ethics as a structural property of governance , not a voluntary moral stance. As consolidated in Table 5 , enforceability is operationalized through bias-linked procurement requirements, mandatory algorithmic audits, accountability clauses, and sanctions for non-compliance. These instruments transform ethics into an observable and evaluable practice, aligning with critiques of “ethics washing” in AI governance (Metcalf et al., 2019; Bietti, 2020 ). In educational contexts—where AI systems directly affect assessment, access, and learner trajectories—structural enforceability is essential for ethical credibility. Multi-Scalar Architecture and Integration FGAIEE operates across interconnected global, institutional, and local scales. Figure 3 visualizes this multi-scalar architecture, demonstrating how global standards and infrastructures interact dynamically with institutional policies and local ethical deliberation. Importantly, no single scale is privileged; ethical alignment emerges through continuous negotiation rather than top-down imposition. By explicitly linking regional disparities to governance mechanisms and outcomes, Fig. 3 advances existing AI ethics models by operationalizing decolonial theory into a multi-scalar governance pathway. This architecture explains why purely global or purely local ethics frameworks are insufficient. FGAIEE instead offers a procedural model that enables contextual articulation of ethics while maintaining global interoperability—an approach particularly suited to education systems characterized by cultural diversity and infrastructural inequality. The figure illustrates FGAIEE as an integrative framework connecting epistemic justice, participatory governance, and structural accountability across global–local educational contexts. Derived from empirical synthesis, the framework demonstrates how ethical principles in AI-enabled education become institutionally legitimate and operational through context-sensitive governance mechanisms. Figure 3 depicts a decolonial governance model that operationalizes glocalized ethics by translating regional disparities (Table 5 ) into strategies for structural redress. At the input stage, AI ethics in education is shaped by tensions between Global North frameworks, local epistemologies, and contextual constraints. Through PICo–SWOT synthesis, moderated by decolonial lenses— coloniality of knowledge, epistemic disobedience, pluriversality —the model identifies asymmetries and openings for reform. These are activated by mediators such as local policy seats, participatory co-design, and culturally responsive teacher training, which ensure epistemic inclusion. The resulting outputs—bias audits, South-led open-source toolkits, and glocalized ethical standards—produce outcomes of structural redress, shared authorship, and democratized educational futures. By tracing this causal pathway, Fig. 3 advances beyond principle-based universalism (Floridi et al., 2018 ; UNESCO, 2021 ) and reframes AI ethics in education as a multi-scalar political project grounded in redistribution and epistemic justice. Positioning FGAIEE as a Theoretical Contribution Synthesizing empirical findings and theory, FGAIEE advances AI ethics in education in three key ways. First, it reframes ethics as an epistemic and governance challenge rather than a checklist of principles. Second, it integrates decolonial theory with institutional design, translating critique into actionable governance. Third, it provides a transferable framework based on process adaptability , not value standardization. Within the structure of this article, FGAIEE functions as the conceptual resolution of RQ3 , offering a coherent explanation of how AI ethics in education can move from symbolic compliance toward structural redress. Rather than proposing another universal model, the framework redefines ethical AI as a shared, participatory, and enforceable project shaped across epistemic and geopolitical boundaries. Limitations and Future Directions This review, while offering a broad synthesis of AI ethics in education, has several limitations. First, its temporal scope (2015–2025/fraction) may not fully capture rapid advances in generative AI since 2022. Future studies should include real-time updates via preprints and continuous mapping. Second, language and geographic biases persist most literature stems from English-speaking, Global North institutions. Expanding language diversity and supporting open-access publishing in the Global South will help decentralize knowledge production. Third, the analysis is based on secondary data, limiting insights into real-world implementation. Future research should include context-rich, empirical case studies, e.g., Ubuntu-based AI curricula in Africa or ethics rooted in gotong royong in Southeast Asia, to better understand culturally grounded practices. These steps are vital for advancing equitable and inclusive AI in education. 5. Conclusion This systematic review demonstrates that AI ethics in education cannot be understood as a neutral or purely technical domain. Rather, it is a contested epistemic field shaped by asymmetries of power, knowledge production, and governance authority. The evidence synthesized across 84 studies reveals that dominant AI ethics frameworks—largely authored in the Global North—continue to reproduce the coloniality of knowledge by privileging Anglophone datasets, Euro-American moral assumptions, and universalist policy templates. In educational contexts, these dynamics translate into algorithmic exclusion, symbolic inclusion, and ethical compliance without structural transformation. At the same time, this review shows that alternative ethical futures are not hypothetical. Across diverse regions, pluriversal practices grounded in Ubuntu ethics, Indigenous data sovereignty, gotong royong governance, and culturally responsive pedagogies demonstrate that ethical AI in education can be co-produced, contextually legitimate, and socially transformative. These initiatives enact what decolonial scholars describe as epistemic disobedience : a refusal to accept imported ethical templates as universal, and an assertion of local knowledge systems as legitimate foundations for governance. The central theoretical contribution of this study is the Framework for Glocalized AI Ethics in Education (FGAIEE) . Unlike principle-based universalist models, FGAIEE operationalizes ethics as a process of structural redress. It integrates epistemic pluriversality, participatory governance, and enforceable policy mechanisms into a coherent multi-scalar model that links global standards with local legitimacy. By embedding ethical commitments into procurement processes, institutional audits, curriculum design, and research funding structures, FGAIEE shifts AI ethics in education from aspirational rhetoric to actionable governance. Methodologically, the study advances a novel PICo–Thematic–SWOT synthesis that demonstrates how critical theory can be translated into policy-relevant insights without sacrificing analytical depth. This approach offers a replicable model for future reviews of ethical AI across sectors, particularly in contexts where global frameworks intersect with local realities. Rather than seeking convergence around abstract principles, the analysis shows that divergence, plurality, and contextual specificity are the true indicators of ethical resilience. Practically, the findings identify concrete leverage points—mandatory bias audits, participatory design mandates, South-led research funding, culturally embedded AI literacy, and frugal technological infrastructures—that can democratize AI governance in education. These measures are not supplementary but essential: without redistributing epistemic authority and institutional power, ethical AI risks becoming an alibi for continued technological expansion rather than a safeguard for justice. In conclusion, this study offers both critique and construction. It exposes how current AI ethics regimes in education reproduce digital colonialism, while simultaneously advancing a globally relevant decolonial framework for transformation. By reframing ethics as shared authorship across epistemologies and scales, the Framework for Glocalized AI Ethics in Education positions ethical AI not as a universal checklist, but as an ongoing political project—one that will determine whose knowledge, values, and futures are encoded into the educational systems of the algorithmic age. Declarations Funding Statement This research received no external funding. Author Contribution Dwi Mariyono (DM) conceptualized the study, led the research design, conducted the systematic review and data synthesis, developed the Framework for Glocalized AI Ethics in Education (FGAIEE), and drafted the original manuscript. DM also served as the corresponding author and coordinated all stages of manuscript development and revision.Muhammad Yunus (MY) contributed to data screening, methodological refinement, and critical review of the empirical findings. MY provided substantive intellectual input to the analysis and interpretation of results and reviewed the manuscript for conceptual clarity and academic rigor.Akmal Nur Alif Hidayatullah (ANAH) supported data extraction and organization, assisted in thematic mapping and verification of findings, and contributed to the refinement of figures, tables, and manuscript formatting. ANAH also participated in manuscript revision and proofreading.All authors reviewed and approved the final manuscript and agreed to be accountable for all aspects of the work. Acknowledgement The authors would like to acknowledge the contributions of the global scholarly community whose work formed the empirical foundation of this review. We are particularly indebted to researchers, educators, and policy scholars whose openly accessible publications enabled a comprehensive and inclusive synthesis across diverse geopolitical and epistemic contexts.We also acknowledge the constructive role of international peer-review standards and open academic infrastructures that support transparent, cumulative, and globally accessible knowledge production in the field of artificial intelligence in education. Data Availability All data underlying this study are derived from publicly accessible and peer-reviewed sources. The reviewed materials consist exclusively of journal articles, reports, and policy documents indexed in internationally recognized academic databases and repositories, including SpringerLink, Scopus, Web of Science, ERIC, and publisher websites of international organizations.No proprietary datasets were used. All sources analyzed in this review are cited in the reference list and can be accessed online without restriction, ensuring full transparency and reproducibility of the review process. References Abdalla, S., Abdeh Kolahchi, A., Ablain, M., Adusumilli, S., Aich Bhowmick, S., Alou-Font, E., Amarouche, L., Andersen, O. B., Antich, H., Aouf, L., Arbic, B., Armitage, T., Arnault, S., Artana, C., Aulicino, G., Ayoub, N., Badulin, S., Baker, S., Banks, C., & Zlotnicki, V. (2021). Altimetry for the future: Building on 25 years of progress. Advances in Space Research , 68 (2), 319–363. https://doi.org/10.1016/j.asr.2021.01.022 Ammah, L. N. A., Lütge, C., Kriebitz, A., & Ramkissoon, L. (2024). AI4people – an ethical framework for a good AI society: the Ghana (Ga) perspective. Journal of Information Communication and Ethics in Society , 22 (4), 453–465. https://doi.org/10.1108/JICES-06-2024-0072 Arora, A., Barrett, M., Lee, E., Oborn, E., & Prince, K. (2023). Risk and the future of AI: Algorithmic bias, data colonialism, and marginalization. Information and Organization , 33 (3), 100478. https://doi.org/10.1016/j.infoandorg.2023.100478 Baker, R. S., & Hawn, A. (2022). Algorithmic Bias in Education. International Journal of Artificial Intelligence in Education , 32 (4), 1052–1092. https://doi.org/10.1007/s40593-021-00285-9 Batool, A., Zowghi, D., & Bano, M. (2025). AI governance: a systematic literature review. AI and Ethics , 5 (3), 3265–3279. https://doi.org/10.1007/s43681-024-00653-w Bietti, E. (2020). From ethics washing to ethics bashing. Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency , 210–219. https://doi.org/10.1145/3351095.3372860 Birhane, A. (2021). Algorithmic injustice: a relational ethics approach. Patterns , 2 (2), 100205. https://doi.org/10.1016/j.patter.2021.100205 Braun, V., & Clarke, V. (2006). Qualitative Research in Psychology Using thematic analysis in psychology Using thematic analysis in psychology. Qualitative Research in Psychology , 3 (2), 77–101. http://www.tandfonline.com/action/journalInformation?journalCode=uqrp20%5Cnhttp://www.tandfonline.com/action/journalInformation?journalCode=uqrp20 Braun, V., & Clarke, V. (2021). Conceptual and Design Thinking for Thematic Analysis. Qualitative Psychology , 9 (1), 3–26. https://doi.org/10.1037/qup0000196 Carlisle, K., & Gruby, R. L. (2019). Polycentric Systems of Governance: A Theoretical Model for the Commons. Policy Studies Journal , 47 (4), 927–952. https://doi.org/10.1111/psj.12212 Corrigan, C. C., Asakipaam, S. A., Kponyo, J. J., & Luetge, C. (Eds.). (2023). AI Ethics in Higher Education: Insights from Africa and Beyond . Springer International Publishing. https://doi.org/10.1007/978-3-031-23035-6 Couldry, N., & Mejias, U. A. (2019). Data Colonialism: Rethinking Big Data’s Relation to the Contemporary Subject. Television & New Media , 20 (4), 336–349. https://doi.org/10.1177/1527476418796632 Davy Tsz Kit, N. G., Luo, W., Chan, H. M. Y., & Chu, S. K. W. (2022). Using digital story writing as a pedagogy to develop AI literacy among primary students. Computers and Education: Artificial Intelligence , 3 , 100054. https://doi.org/10.1016/j.caeai.2022.100054 de Sousa Santos, B. (2023). Epistemologies of the south: Justice against epistemicide. In Foundational Concepts of Decolonial and Southern Epistemologies . https://doi.org/10.1177/0094306115609925nn du Boulay, B. (2023). Artificial Intelligence in Education and Ethics. In Handbook of Open, Distance and Digital Education (pp. 93–108). Springer Nature Singapore. https://doi.org/10.1007/978-981-19-2080-6_6 Floridi, L., Cowls, J., Beltrametti, M., Chatila, R., Chazerand, P., Dignum, V., Luetge, C., Madelin, R., Pagallo, U., Rossi, F., Schafer, B., Valcke, P., & Vayena, E. (2018). AI4People—An Ethical Framework for a Good AI Society: Opportunities, Risks, Principles, and Recommendations. Minds and Machines , 28 (4), 689–707. https://doi.org/10.1007/s11023-018-9482-5 Heeks, R. (2022). Digital inequality beyond the digital divide: conceptualizing adverse digital incorporation in the global South. Information Technology for Development , 28 (4), 688–704. https://doi.org/10.1080/02681102.2022.2068492 Helms, M. M., & Nixon, J. (2010). Exploring SWOT analysis – where are we now? A review of academic research from the last decade. Journal of Strategy and Management , 3 (3), 215–251. https://doi.org/10.1108/17554251011064837 Holmes, W., Porayska-Pomsta, K., Holstein, K., Sutherland, E., Baker, T., Shum, S. B., Santos, O. C., Rodrigo, M. T., Cukurova, M., Bittencourt, I. I., & Koedinger, K. R. (2022). Ethics of AI in Education: Towards a Community-Wide Framework. International Journal of Artificial Intelligence in Education , 32 (3), 504–526. https://doi.org/10.1007/s40593-021-00239-1 Jobin, A., Ienca, M., & Vayena, E. (2019). The global landscape of AI ethics guidelines. Nature Machine Intelligence , 1 (9), 389–399. https://doi.org/10.1038/s42256-019-0088-2 Kamali, J. (2024). AI ethics as a complex and multifaceted challenge: decoding educators’ AI ethics alignment through the lens of activity theory. International Journal of Educational Technology in Higher Education , 21 (1). https://doi.org/10.1186/s41239-024-00496-9 Khan, A. A., Badshah, S., Liang, P., Waseem, M., Khan, B., Ahmad, A., Fahmideh, M., Niazi, M., & Akbar, M. A. (2022). Ethics of AI: A Systematic Literature Review of Principles and Challenges. The International Conference on Evaluation and Assessment in Software Engineering 2022 , 383–392. https://doi.org/10.1145/3530019.3531329 Koole, M., Smith, M., Traxler, J., Adam, T., & Footring, S. (2024). Decolonising Educational Technology. Education Sciences , 14 (10), 1070. https://doi.org/10.3390/educsci14101070 Kwet, M. (2019). Digital colonialism: US empire and the new imperialism in the Global South. Race & Class , 60 (4), 3–26. https://doi.org/10.1177/0306396818823172 Madlberger, L. (2017). Gotong royong in NLP research a mobile tool for collaborative text annotation in Indonesia. In Proceedings of the 2016 International Conference on Asian Language Processing Ialp 2016 (pp. 99–102). https://doi.org/10.1109/IALP.2016.7875944 McHugh, M. L. (2012). Interrater reliability: the kappa statistic. Biochemia Medica , 276–282. https://doi.org/10.11613/BM.2012.031 Mehrabi, N., Morstatter, F., Saxena, N., Lerman, K., & Galstyan, A. (2022). A Survey on Bias and Fairness in Machine Learning. ACM Computing Surveys , 54 (6), 1–35. https://doi.org/10.1145/3457607 Mignolo, W. D. (2020). The Darker Side of Western Modernity. The Darker Side of Western Modernity . Duke University Press. https://doi.org/10.2307/j.ctv125jqbw Mohamed, S., Png, M. T., & Isaac, W. (2020). Decolonial AI: Decolonial Theory as Sociotechnical Foresight in Artificial Intelligence. Philosophy & Technology , 33 (4), 659–684. https://doi.org/10.1007/s13347-020-00405-8 Nazer, L. H., Zatarah, R., Waldrip, S., Ke, J. X. C., Moukheiber, M., Khanna, A. K., Hicklen, R. S., Moukheiber, L., Moukheiber, D., Ma, H., & Mathur, P. (2023). Bias in artificial intelligence algorithms and recommendations for mitigation. PLOS Digital Health , 2 (6), e0000278. https://doi.org/10.1371/journal.pdig.0000278 Newman, S. J. (2024). Aníbal Quijano’s Critical Sociology: From Dependency Theory to Coloniality. Sociology of Race and Ethnicity , 10 (4), 566–571. https://doi.org/10.1177/23326492241268532 Ng, D. T. K., Leung, J. K. L., Chu, K. W. S., & Qiao, M. S. (2021a). AI Literacy: Definition, Teaching, Evaluation and Ethical Issues. Proceedings of the Association for Information Science and Technology , 58 (1), 504–509. https://doi.org/10.1002/pra2.487 Ng, D. T. K., Leung, J. K. L., Chu, S. K. W., & Qiao, M. S. (2021b). Conceptualizing AI literacy: An exploratory review. Computers and Education: Artificial Intelligence , 2 , 100041. https://doi.org/10.1016/j.caeai.2021.100041 Nurhafida, Y. H., Muchtarom, M., & Gunawati, D. (2020). Strategi FKPM (Forum Kemitraan Polri dan Masyarakat) dalam Meningkatkan Kesadaran Hukum Masyarakat (Studi di Kelurahan Kadipiro Kecamatan Banjarsari Kota Surakarta). Jurnal PPKn , 8 (1), 1–8. https://doi.org/https://doi.org/10.2019/jppkn.v8i1.73 O’cathain, A., Murphy, E., & Nicholl, J. (2008). The Quality of Mixed Methods Studies in Health Services Research. Journal of Health Services Research & Policy , 13 (2), 92–98. https://doi.org/10.1258/jhsrp.2007.007074 OECD (2023). OECD Digital Education Outlook 2023 . OECD. https://doi.org/10.1787/589b283f-en Olojede, H. T. (2023). Towards African Artificial Intelligence Ethical Principles. 2023 First International Conference on the Advancements of Artificial Intelligence in African Context (AAIAC) , 1–6. https://doi.org/10.1109/AAIAC60008.2023.10465379 Ostrom, E. (1990). Governing the commons: the evolution of institutions for collective action. In Governing the commons: the evolution of institutions for collective action . https://doi.org/10.2307/3146384 Ostrom, E. (2010). Beyond Markets and States: Polycentric Governance of Complex Economic Systems. American Economic Review , 100 (3), 641–672. https://doi.org/10.1257/aer.100.3.641 Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., Glanville, J., Grimshaw, J. M., Hróbjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-Wilson, E., McDonald, S., & Moher, D. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. International Journal of Surgery , 88 , 105906. https://doi.org/10.1016/j.ijsu.2021.105906 Papakostas, C. (2025). Artificial Intelligence in Religious Education: Ethical, Pedagogical, and Theological Perspectives. Religions , 16 (5), 563. https://doi.org/10.3390/rel16050563 Papyshev, G., & Chan, K. J. D. (2025). Fugazi regulation for AI: strategic tolerance for ethics washing. AI & SOCIETY , 40 (5), 3475–3486. https://doi.org/10.1007/s00146-024-02084-x Pasupuleti, M. K. (2024). Ethical AI Governance: A Global Blueprint. In AI Governance: Innovating with Integrity (pp. 60–77). National Education Services. https://doi.org/10.62311/nesx/97832 Quijano, A. (2000). Coloniality of Power and Eurocentrism in Latin America. International Sociology , 15 (2), 215–232. https://doi.org/10.1177/0268580900015002005 Sahebi, S., & Formosa, P. (2024). Artificial Intelligence (AI) and Global Justice. Minds and Machines , 35 (1), 4. https://doi.org/10.1007/s11023-024-09708-7 Shams, R. A., Zowghi, D., & Bano, M. (2023). AI and the quest for diversity and inclusion: a systematic literature review. AI and Ethics . https://doi.org/10.1007/s43681-023-00362-w Stern, C., Jordan, Z., & Mcarthur, A. (2014). Developing the review question and inclusion criteria. American Journal of Nursing , 114 (4), 53–56. https://doi.org/10.1097/01.NAJ.0000445689.67800.86 UNESCO. (2021). AI and education: guidance for policy-makers. In AI and education: guidance for policy-makers . UNESCO Publishing . https://doi.org/10.54675/pcsp7350 Upadhyay, U., Gradisek, A., Iqbal, U., Dhar, E., Li, Y. C., & Syed-Abdul, S. (2023). Call for the responsible artificial intelligence in the healthcare. BMJ Health & Care Informatics , 30 (1), e100920. https://doi.org/10.1136/bmjhci-2023-100920 Vetter, M. A., Lucia, B., Jiang, J., & Othman, M. (2024). Towards a framework for local interrogation of AI ethics: A case study on text generators, academic integrity, and composing with ChatGPT. Computers and Composition , 71 , 102831. https://doi.org/10.1016/j.compcom.2024.102831 Wakunuma, K., & Eke, D. (2024). Africa, ChatGPT, and Generative AI Systems: Ethical Benefits, Concerns, and the Need for Governance. Philosophies , 9 (3). https://doi.org/10.3390/philosophies9030080 Wiese (2025). AI Ethics in Education Systematic Literature Review . Grail (Governance & Responsible AI LAB). https://www.grail-lab.org/publications/ai-ethics-in-education-systematic-literature-review?utm_source=chatgpt.com Yadav, R., Yadav, M., Huzooree, G., Najeeb, A., & Ranasinghe, A. (2025). Accessibility and inclusivity in AI-powered personalized learning. Transformative AI Practices for Personalized Learning Strategies , 47–67. https://doi.org/10.4018/979-8-3693-8744-3.ch003 Yilma, K. (2025). Ethics of AI in Africa: Interrogating the role of Ubuntu and AI governance initiatives. Ethics and Information Technology , 27 (2), 24. https://doi.org/10.1007/s10676-025-09834-5 Zajko, M. (2021). Conservative AI and social inequality: conceptualizing alternatives to bias through social theory. AI & SOCIETY , 36 (3), 1047–1056. https://doi.org/10.1007/s00146-021-01153-9 Zembylas, M. (2023). A decolonial approach to AI in higher education teaching and learning: strategies for undoing the ethics of digital neocolonialism. Learning Media and Technology , 48 (1), 25–37. https://doi.org/10.1080/17439884.2021.2010094 Additional Declarations No competing interests reported. Supplementary Files Appendix.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8417921","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":564492963,"identity":"52c6ef2a-ebf4-4202-818f-18d4b968b656","order_by":0,"name":"Dwi Mariyono","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA90lEQVRIiWNgGAWjYDCCA4wNQFICxGR8AGWAgARYnJAWZgMitSCYbBLIEji18N1Ibt3wc49FYr9087Fq3h0W9rrtBxg//GCwkMWlRfJGYtvNnmcSiTPnHEu7zXtGInHbmQRmyR4GCWNcWgxuJ7bd4DkgkbjhRo7Zbd42iQSzAwkM0kC/JOLTcvMPVEsxUIu92fkHzL8JabkNs4UZqIVx240ENry2SN5/2HZb5oCE8cwZacmSc9uAfrnxsM2yxwC3X/jOHH92882BOtl+ieSDH9621QEdlnz4xo+KOpwhBgOOSApAMWJAQD0Q2BNWMgpGwSgYBSMWAADXnV8k5zwDpAAAAABJRU5ErkJggg==","orcid":"","institution":"Universitas Islam Malang","correspondingAuthor":true,"prefix":"","firstName":"Dwi","middleName":"","lastName":"Mariyono","suffix":""},{"id":564492964,"identity":"723a9342-f232-4497-ac6d-e983a819ff8a","order_by":1,"name":"Muhammad Yunus","email":"","orcid":"","institution":"Universitas Islam Malang","correspondingAuthor":false,"prefix":"","firstName":"Muhammad","middleName":"","lastName":"Yunus","suffix":""},{"id":564492965,"identity":"5760d109-58d3-438d-8d71-b721f33f69a8","order_by":2,"name":"Akmal Nur Alif Hidayatullah","email":"","orcid":"","institution":"University of Brawijaya","correspondingAuthor":false,"prefix":"","firstName":"Akmal","middleName":"Nur Alif","lastName":"Hidayatullah","suffix":""}],"badges":[],"createdAt":"2025-12-21 14:38:39","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-8417921/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8417921/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":98972898,"identity":"e2400c8e-7b25-40a3-b825-ff559da95beb","added_by":"auto","created_at":"2025-12-25 03:08:22","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":328789,"visible":true,"origin":"","legend":"","description":"","filename":"DecolonizingAIEthicsinEducation20122025.docx","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/1523fa22dd4c0e1d44471a19.docx"},{"id":98972896,"identity":"478ff056-e820-4f40-91fe-11b195f3c802","added_by":"auto","created_at":"2025-12-25 03:08:15","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7925,"visible":true,"origin":"","legend":"","description":"","filename":"1ae614c0f2074fbfab359ba97a34a983.json","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/81e3ff3364d982f2b351f2fc.json"},{"id":98972895,"identity":"a0d4f72f-9b2b-4bb7-9e1b-ea04bcce471f","added_by":"auto","created_at":"2025-12-25 03:08:15","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":122496,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/32cf1b708a2274e13320a157.docx"},{"id":98972893,"identity":"3d360bd5-176f-450f-aa01-14f82a827883","added_by":"auto","created_at":"2025-12-25 03:08:15","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":170542,"visible":true,"origin":"","legend":"","description":"","filename":"1ae614c0f2074fbfab359ba97a34a9831enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/be3cb306196f3c40be1f1a9c.xml"},{"id":98972892,"identity":"9031274b-5aa6-4c7d-b1b4-444db84cbf30","added_by":"auto","created_at":"2025-12-25 03:08:14","extension":"jpeg","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1074,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/20a0c31f4f3cec303341ce65.jpeg"},{"id":98972897,"identity":"c767c3ec-eeb9-4645-9abb-70da5dc63670","added_by":"auto","created_at":"2025-12-25 03:08:22","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":111868,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/3263a7360223dbe2080ce861.png"},{"id":98972941,"identity":"13996bee-52e9-424b-a5e6-3d5deb9f91f8","added_by":"auto","created_at":"2025-12-25 03:08:26","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":84006,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/94b059a115e3c008dba02d9b.png"},{"id":98972900,"identity":"3b007f8b-b707-44a0-a54a-436bae23ed79","added_by":"auto","created_at":"2025-12-25 03:08:22","extension":"jpeg","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":102613,"visible":true,"origin":"","legend":"","description":"","filename":"groupimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/fcf03c1c80054007c13db4a6.jpeg"},{"id":98972901,"identity":"26f655d8-b67c-4ae2-ab64-1d0b28904e36","added_by":"auto","created_at":"2025-12-25 03:08:23","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":935,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/bb9e2578269e0f6040ce576b.png"},{"id":98972904,"identity":"0fd5859e-6e4f-45a3-adc9-c7bf788b63bd","added_by":"auto","created_at":"2025-12-25 03:08:23","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":34410,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/cadefe5a7df50d2446a5474d.png"},{"id":98972925,"identity":"d6fe3843-b09a-43a7-9b51-4f0fa92d7413","added_by":"auto","created_at":"2025-12-25 03:08:24","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":27445,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/3640f850d40b149fcb4f831f.png"},{"id":98972922,"identity":"5de3fb59-5869-4ebd-ad48-8ac6f4cb3299","added_by":"auto","created_at":"2025-12-25 03:08:24","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":27140,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinegroupimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/4db33cc1694c32de89ccafeb.png"},{"id":98972902,"identity":"0aece6c9-a150-4389-b2ac-2deab45dfb37","added_by":"auto","created_at":"2025-12-25 03:08:23","extension":"xml","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":168166,"visible":true,"origin":"","legend":"","description":"","filename":"1ae614c0f2074fbfab359ba97a34a9831structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/120f2879de0c9d7ac14e8a59.xml"},{"id":98972923,"identity":"c52e9ef8-1361-41b9-9ba0-b17be011ec4f","added_by":"auto","created_at":"2025-12-25 03:08:24","extension":"html","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":186473,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/9b933fef5930c417aa772ab5.html"},{"id":99312339,"identity":"71c85e5b-2e13-40c3-adf5-a8e6ff3a023a","added_by":"auto","created_at":"2025-12-31 16:18:50","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":46071,"visible":true,"origin":"","legend":"\u003cp\u003ePRISMA 2020 Flow Diagram for AI Ethics in Education Review\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/d63372ecb0ea9df438b9f86d.png"},{"id":98972894,"identity":"61883fd1-95d8-4058-9b4d-e23e31a08e8d","added_by":"auto","created_at":"2025-12-25 03:08:15","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":170074,"visible":true,"origin":"","legend":"\u003cp\u003eFrom Symbolic Ethical Compliance to Structural Redress in AI Governance for Education.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/4716c7996810e9950f1994b6.png"},{"id":98972909,"identity":"fdc25ea5-d98e-4e50-a589-d2133d288f9a","added_by":"auto","created_at":"2025-12-25 03:08:23","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":165704,"visible":true,"origin":"","legend":"\u003cp\u003eThe Framework for Glocalized AI Ethics in Education (FGAIEE).\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/683f2c530b2be2b0430276fd.png"},{"id":104958316,"identity":"22e744b1-7655-4d1e-ab60-4ffe44895140","added_by":"auto","created_at":"2026-03-19 08:27:53","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2111064,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/7b220b0d-23a9-4442-83d2-627e0e1d1722.pdf"},{"id":98972906,"identity":"ad0c9487-ffd1-4da7-bca0-e6bd7319884b","added_by":"auto","created_at":"2025-12-25 03:08:23","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":122496,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-8417921/v1/7a8456a2be7a850c8f214cb5.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Decolonizing AI Ethics in Education: A Systematic Review and the Framework for Glocalized AI Ethics in Education (FGAIEE)","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eArtificial Intelligence (AI) is rapidly reshaping educational systems worldwide, influencing pedagogical practices, assessment regimes, institutional governance, and policy infrastructures. Framed as a catalyst for personalization, efficiency, and scalability, AI has become embedded in global education reforms\u0026mdash;from automated grading and adaptive learning systems to predictive analytics and administrative decision-making. Yet, as AI technologies proliferate across educational contexts, the ethical frameworks governing their design and deployment remain overwhelmingly shaped by Euro-American epistemologies and Global North policy imaginaries. This imbalance raises a critical question: \u003cem\u003eethical for whom, by whom, and under what epistemic authority\u003c/em\u003e?\u003c/p\u003e \u003cp\u003eContemporary AI ethics in education is dominated by universalist frameworks produced by international organizations and Global North institutions, such as the European Union\u0026rsquo;s Ethics Guidelines for Trustworthy AI, UNESCO\u0026rsquo;s AI and Education recommendations, and the OECD AI Principles (Floridi et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Jobin et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; OECD, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; UNESCO, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). While these initiatives have significantly advanced global awareness of fairness, transparency, and accountability, they are typically formulated within contexts of robust digital infrastructure, Anglophone datasets, and liberal-individualist moral traditions. When exported as \u0026ldquo;global standards,\u0026rdquo; these frameworks often fail to account for the cultural, linguistic, political, and infrastructural realities of the Global South, thereby reproducing what critical scholars describe as \u003cem\u003edigital colonialism\u003c/em\u003e (Arora et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Couldry \u0026amp; Mejias, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Kwet, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eExisting frameworks such as the community-wide ethics model for AIED (du Boulay, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Holmes et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) and studies on AI ethics education (Wiese et al., \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2025\u003c/span\u003e) provide foundational principles for fairness and accountability in educational AI, but they remain largely descriptive and insufficiently connected to governance structures that account for epistemic asymmetries across diverse sociocultural contexts.\u003c/p\u003e \u003cp\u003eIn educational settings, the consequences of this imbalance are increasingly evident. AI-driven assessment systems systematically underperform for non-native English speakers; algorithmic decision-making privileges dominant cultural norms embedded in training data; and governance mechanisms rarely provide meaningful participatory space for educators, learners, or policymakers from marginalized regions (Baker \u0026amp; Hawn, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Heeks, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Zajko, \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Ethical safeguards designed to promote equity thus risk entrenching inequality\u0026mdash;transforming AI ethics from a protective instrument into a vehicle of epistemic exclusion. This phenomenon reveals a deeper problem: AI ethics in education is not merely a technical or normative domain, but a \u003cem\u003esite of epistemic power and geopolitical struggle\u003c/em\u003e.\u003c/p\u003e \u003cp\u003eDespite a growing body of scholarship on AI ethics and digital colonialism, three critical gaps persist. \u003cb\u003eFirst\u003c/b\u003e, existing AI ethics literature overwhelmingly treats ethics as a universal, value-neutral checklist, insufficiently interrogating whose epistemologies define ethical legitimacy and how power asymmetries shape \u0026ldquo;global\u0026rdquo; standards ((Floridi et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Jobin et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). \u003cb\u003eSecond\u003c/b\u003e, systematic reviews in this field tend to focus either on technical fairness metrics (Mehrabi et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) or on high-level mappings of ethical principles, without offering integrative methodological approaches capable of translating critical insights into actionable governance mechanisms. \u003cb\u003eThird\u003c/b\u003e, global AI ethics frameworks in education remain largely principle-based, resulting in symbolic compliance rather than enforceable structural redress\u0026mdash;particularly in contexts marked by infrastructural constraints and epistemic marginalization (Sahebi \u0026amp; Formosa, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eTo address these gaps, this study reconceptualizes AI ethics in education through the lenses of \u003cem\u003eepistemic pluriversality\u003c/em\u003e and \u003cem\u003eglocalized governance\u003c/em\u003e. Drawing on decolonial theory, pluriversality recognizes multiple knowledge traditions\u0026mdash;such as Ubuntu ethics, Indigenous data sovereignty, Islamic epistemologies, Confucian relational ethics, and \u003cem\u003egotong royong\u003c/em\u003e governance\u0026mdash;not as cultural add-ons, but as coequal sources of ethical authority (Mignolo, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Nurhafida et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Zembylas, \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Glocalized governance, in turn, rejects the unidirectional export of global norms and instead embeds ethical principles within locally determined practices, ensuring that AI systems are co-designed, audited, and governed by the communities they affect (Abdalla et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Olojede, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; UNESCO, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eMethodologically, this study conducts a PRISMA-guided systematic review of 84 peer-reviewed studies published between 2015 and 2025, employing an original \u003cb\u003ePICo\u0026ndash;Thematic\u0026ndash;SWOT synthesis\u003c/b\u003e. This multi-layered approach integrates: (1) the PICo framework to foreground populations, interests, and contexts that are often marginalized; (2) thematic analysis to surface epistemic patterns and ethical tensions; and (3) SWOT analysis as an epistemic diagnostic tool that translates critical findings into policy-relevant leverage points. Unlike conventional reviews, this design moves beyond description to generate strategic pathways for ethical governance in education.\u003c/p\u003e \u003cp\u003eThe study is guided by three research questions:\u003c/p\u003e \u003cp\u003e \u003cb\u003eRQ1 (Diagnostic \u0026ndash; Coloniality as Mechanism)\u003c/b\u003e:\u003c/p\u003e \u003cp\u003e \u003cem\u003eHow do dominant AI ethics frameworks in education operationalize assumptions of universal ethical transferability, and through what sociotechnical mechanisms\u0026mdash;such as datasets, procurement regimes, benchmarking practices, and accreditation systems\u0026mdash;do they reproduce geopolitical and epistemic asymmetries between the Global North and the Global South?\u003c/em\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eRQ2 (Analytic \u0026ndash; Epistemic Disobedience \u0026amp; Pluriversality)\u003c/b\u003e:\u003c/p\u003e \u003cp\u003e \u003cem\u003eWhat forms of epistemic disobedience and pluriversal ethical practices have emerged within educational AI systems across diverse contexts, and how do these practices reconfigure participation, accountability, and ethical legitimacy beyond universalist frameworks?\u003c/em\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eRQ3 (Constructive \u0026ndash; From Principles to Governance)\u003c/b\u003e \u003c/p\u003e \u003cp\u003eHow can AI ethics in education be reconceptualized from principle-based or symbolic compliance toward structurally inclusive and glocalized governance models that institutionalize epistemic pluriversality across policy, design, and implementation?\u003c/p\u003e \u003cp\u003eIn response to these questions, the study advances its central contribution: the \u003cb\u003eFramework for Glocalized AI Ethics in Education (FGAIEE)\u003c/b\u003e. FGAIEE reconceptualizes ethical AI governance as a multi-scalar, participatory process that bridges global infrastructures with local epistemologies through enforceable mechanisms such as bias-linked procurement, participatory audits, South-led open-source toolkits, and culturally responsive teacher training. In doing so, this article positions AI ethics in education not as a universal moral template, but as a political project of epistemic justice and shared authorship\u0026mdash;one that will ultimately shape whether AI entrenches or transforms global educational futures. Rather than presenting a separate theoretical framework section, this study integrates critical and decolonial perspectives into the introduction and develops them through empirical synthesis and theoretical construction in the Discussion.\u003c/p\u003e"},{"header":"2. Method","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Research Design\u003c/h2\u003e \u003cp\u003eThis study adopts a PRISMA-guided systematic literature review to critically examine AI ethics in education through a decolonial and governance-oriented lens (Page et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Unlike conventional AI ethics reviews that primarily synthesize ethical principles or normative positions (Floridi et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Jobin et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), this review is designed to interrogate how \u003cem\u003eethical authority, epistemic power, and governance logics\u003c/em\u003e are constructed, circulated, and contested across global educational contexts.\u003c/p\u003e \u003cp\u003eTo address this epistemically complex problem, the study employs a layered methodological architecture integrating PICo framing, thematic analysis, and SWOT synthesis. This design reflects calls within critical AI studies to move beyond descriptive aggregation toward explanatory and governance-relevant synthesis (Heeks, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The methodological sequence allows the analysis to progress from contextual specification (PICo), to interpretive pattern detection (thematic analysis), and finally to policy-operable synthesis (SWOT).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Review Protocol and Search Strategy\u003c/h2\u003e \u003cp\u003eThe review followed the PRISMA 2020 statement to ensure transparency, rigor, and replicability in study identification, screening, and inclusion (Page et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Searches were conducted across Scopus, Web of Science, ERIC, and Google Scholar, which collectively capture dominant, interdisciplinary, and education-focused scholarship.\u003c/p\u003e \u003cp\u003eSearch strings combined terms related to \u003cem\u003eartificial intelligence\u003c/em\u003e, \u003cem\u003ealgorithmic systems\u003c/em\u003e, \u003cem\u003eethics\u003c/em\u003e, \u003cem\u003egovernance\u003c/em\u003e, \u003cem\u003eeducation\u003c/em\u003e, \u003cem\u003eequity\u003c/em\u003e, \u003cem\u003eGlobal South\u003c/em\u003e, and \u003cem\u003edecolonial perspectives\u003c/em\u003e, using Boolean operators to balance breadth and specificity. In line with critiques of epistemic exclusion in AI research (Couldry \u0026amp; Mejias, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Mohamed et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), the search strategy intentionally avoided limiting results to formal AI ethics frameworks, enabling the inclusion of studies addressing Indigenous data sovereignty, pluriversal ethics, and localized governance practices.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Inclusion and Exclusion Criteria\u003c/h2\u003e \u003cp\u003eStudies were included if they:\u003c/p\u003e \u003cp\u003e \u003col\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eExamined AI or algorithmic systems in educational contexts (K\u0026ndash;12, higher education, lifelong learning, or digital learning platforms);\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eEngaged explicitly with ethical, governance, justice, or power-related issues;\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eWere peer-reviewed journal articles published between 2015 and 2025, reflecting the maturation of AI ethics discourse; and\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eProvided empirical, theoretical, or policy-relevant insights.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003c/ol\u003e \u003c/p\u003e \u003cp\u003eStudies were excluded if they:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eFocused solely on technical optimization without ethical or governance analysis (Mehrabi et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2022\u003c/span\u003e);\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eAddressed AI ethics outside education;\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eWere editorial commentaries without analytical grounding; or\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eLacked sufficient methodological transparency.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eFollowing screening and eligibility assessment, \u003cb\u003e84 studies\u003c/b\u003e were retained for final analysis \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 PICo Framework for Analytical Structuring\u003c/h2\u003e \u003cp\u003eThe \u003cb\u003ePICo framework\u003c/b\u003e\u0026mdash;Population, Interest, and Context\u0026mdash;was employed to structure the review and maintain analytical coherence across heterogeneous studies (Stern et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003ePopulation (P)\u003c/b\u003e: Educational actors and systems, including learners, educators, institutions, and policymakers across Global North and Global South contexts.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eInterest (I)\u003c/b\u003e: Ethical governance of AI in education, with emphasis on epistemic authority, algorithmic bias, participation, and accountability.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eContext (Co)\u003c/b\u003e: Sociotechnical, cultural, and geopolitical settings in which AI ethics frameworks are developed, implemented, and legitimized.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003ePICo was used not merely as a categorization tool, but as an \u003cem\u003einterpretive scaffold\u003c/em\u003e that foregrounds context as analytically central\u0026mdash;responding to critiques that ethics frameworks often treat educational settings as neutral implementation spaces (Baker \u0026amp; Hawn, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Zajko, \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Zembylas, \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Data Extraction and Thematic Analysis\u003c/h2\u003e \u003cp\u003eData extraction captured publication characteristics, ethical frameworks referenced, educational contexts, governance mechanisms, and reported ethical tensions or outcomes. A \u003cem\u003ehybrid inductive\u0026ndash;deductive thematic analysis\u003c/em\u003e was conducted following the reflexive approach outlined by Braun and Clarke (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2006\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eDeductive coding was informed by critical AI ethics and decolonial scholarship, including concepts such as \u003cem\u003euniversality\u003c/em\u003e, \u003cem\u003eepistemic justice\u003c/em\u003e, \u003cem\u003edigital colonialism\u003c/em\u003e, and \u003cem\u003eparticipatory governance\u003c/em\u003e (Birhane, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Kwet, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Mignolo, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Inductive coding allowed unanticipated patterns\u0026mdash;particularly counterintuitive and tension findings\u0026mdash;to emerge from the data. This approach ensured theoretical sensitivity without constraining analytical openness.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.6 Inter-Rater Reliability and Analytical Rigor\u003c/h2\u003e \u003cp\u003eTo enhance analytical rigor and credibility, coding was conducted independently by multiple reviewers. Inter-rater reliability was assessed using Cohen\u0026rsquo;s kappa, (Appendix Table\u0026nbsp;3B) yielding agreement levels categorized as \u003cem\u003eexcellent\u003c/em\u003e according to established benchmarks (McHugh, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Discrepancies were resolved through iterative dialogue, prioritizing interpretive alignment over mechanical consensus.\u003c/p\u003e \u003cp\u003eSuch procedures are particularly important in reviews engaging with normative and epistemic questions, where analytical transparency is essential to scholarly trustworthiness (O\u0026rsquo;cathain et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2008\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.7 SWOT Synthesis for Governance-Oriented Interpretation\u003c/h2\u003e \u003cp\u003eWhile thematic analysis captured epistemic patterns and power relations, a further analytical step was required to translate critical insights into \u003cb\u003egovernance-relevant implications\u003c/b\u003e. To this end, a \u003cb\u003eSWOT synthesis\u003c/b\u003e was employed at the interpretive stage.\u003c/p\u003e \u003cp\u003eConsistent with prior critical policy reviews (Heeks, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Helms \u0026amp; Nixon, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2010\u003c/span\u003e), SWOT was repurposed not as a managerial tool but as a \u003cb\u003estructuring device\u003c/b\u003e to connect epistemic critique with policy operability. Strengths and weaknesses reflected internal characteristics of dominant AI ethics frameworks, while opportunities and threats captured external geopolitical, institutional, and cultural dynamics shaping educational AI governance.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e2.8 Methodological Justification and Reflexivity\u003c/h2\u003e \u003cp\u003eThe integration of \u003cem\u003ePICo framing, PRISMA-guided screening, thematic analysis, and SWOT synthesis\u003c/em\u003e is not methodological excess, but a deliberate response to a recognized limitation in existing AI ethics reviews. Much of the literature remains confined to principle-level comparison or descriptive synthesis, offering limited insight into how ethical authority is operationalized within real governance infrastructures (Birhane, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Floridi et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Jobin et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eMethodological simplification would risk reproducing the very universalist assumptions this review seeks to interrogate. By contrast, methodological plurality enables the study to bridge \u003cem\u003eepistemic critique with policy-operable governance design\u003c/em\u003e, addressing a gap widely acknowledged yet rarely resolved in AI ethics and education scholarship (Couldry \u0026amp; Mejias, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Heeks, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Mohamed et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cp\u003eThis section reports the empirical findings of the systematic review of 84 peer-reviewed studies (2015\u0026ndash;2025). Results are organized according to the three research questions (RQ1\u0026ndash;RQ3) and present observed patterns, distributions, and documented practices derived from the data. Interpretive, theoretical, and normative analysis is reserved exclusively for the Discussion section.\u003c/p\u003e \u003cp\u003eTwo integrative results artifacts anchor this section:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e2\u003c/span\u003e, which synthesizes dominant empirical themes, representative studies, and observed policy implications; and\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e3\u003c/span\u003e, which maps distributional patterns of AI ethics in education through a decolonial analytical lens.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eTogether, these tables serve distinct but complementary empirical functions: Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e2\u003c/span\u003e synthesizes recurring thematic findings across the literature, while Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e3\u003c/span\u003e maps how these themes are unevenly distributed across geopolitical contexts, datasets, and governance arrangements.\u003c/p\u003e \u003cp\u003e \u003cb\u003e3.1. RQ1. How do current global AI ethics frameworks in education reproduce or challenge colonial epistemologies and structural asymmetries?\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe review reveals a pronounced \u003cem\u003egeopolitical concentration in the production and circulation of AI ethics frameworks applied to education\u003c/em\u003e. Across the dataset, \u003cb\u003e78%\u003c/b\u003e \u003cem\u003eof identified frameworks and policy instruments originated from Global North institutions\u003c/em\u003e, primarily Europe and North America (Khan et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Pasupuleti, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Upadhyay et al., \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). These frameworks were most often implemented in Global South contexts through policy transfer, donor requirements, or procurement conditions, with limited evidence of local co-authorship.\u003c/p\u003e \u003cp\u003eEmpirical studies further document \u003cem\u003esystematic algorithmic performance disparities\u003c/em\u003e. AI-driven assessment and language-processing systems exhibited \u003cem\u003eaccuracy gaps of approximately 10\u0026ndash;15% for non-native English speakers\u003c/em\u003e, compared with native speakers, across multiple educational settings (Baker \u0026amp; Hawn, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Zajko, \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). These disparities were consistently associated with English-dominant training datasets and culturally narrow pedagogical assumptions.\u003c/p\u003e \u003cp\u003eGovernance and participation patterns mirrored these asymmetries. Fewer than \u003cb\u003eone in five studies\u003c/b\u003e reported mechanisms enabling educators, learners, or policymakers from marginalized regions to participate meaningfully in ethical decision-making processes related to AI adoption (Arora et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Sahebi \u0026amp; Formosa, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Where participation was noted, it was predominantly advisory rather than deliberative or decision-making in nature.\u003c/p\u003e \u003cp\u003eThese empirical patterns are synthesized thematically in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e2\u003c/span\u003e, which consolidates recurrent findings related to Global North dominance, algorithmic exclusion, and participatory gaps across studies. Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e3\u003c/span\u003e, by contrast, maps the \u003cem\u003edistributional structure\u003c/em\u003e of these findings, demonstrating how authorship, datasets, and governance authority are disproportionately concentrated in Global North contexts.\u003c/p\u003e \u003cp\u003e \u003cb\u003e3.2. RQ2. What pluriversal counter-strategies and localized practices have emerged to foster epistemic inclusion and ethical legitimacy?\u003c/b\u003e \u003c/p\u003e \u003cp\u003eAlongside dominant global frameworks, the review identified a growing set of \u003cem\u003econtext-specific ethical practices\u003c/em\u003e emerging from Global South, Indigenous, and localized educational settings. These practices were characterized by participatory governance, cultural embeddedness, and local accountability mechanisms.\u003c/p\u003e \u003cp\u003eSeveral studies documented \u003cem\u003eUbuntu-informed co-design models\u003c/em\u003e in African educational contexts. In Ghana, participatory ethical deliberation grounded in communal decision-making was associated with a reported 40% increase in teacher engagement and trust toward AI-supported educational tools (Ammah et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Comparable participatory models rooted in \u003cem\u003egotong royong\u003c/em\u003e principles were reported in Southeast Asian contexts, emphasizing collective responsibility for AI oversight and evaluation (Madlberger, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2017\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eAnother recurring pattern involved \u003cem\u003eIndigenous data sovereignty and custodial governance arrangements\u003c/em\u003e. Studies reported models in which data ownership, consent, and usage were negotiated collectively, positioning educational data as a shared resource rather than an extractable commodity (Zembylas, \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). These approaches were frequently linked to heightened transparency and community trust.\u003c/p\u003e \u003cp\u003eAdditionally, the review identified \u003cem\u003eSouth-led open-source and frugal AI initiatives\u003c/em\u003e as operational counter-strategies to infrastructural dependency. These initiatives prioritized multilingual design, low-bandwidth functionality, and local capacity-building, enabling ethical adaptation in resource-constrained environments (Ng et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2021b\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2021a\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThese pluriversal practices are consolidated empirically in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e2\u003c/span\u003e as recurring thematic patterns with documented outcomes. Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e3\u003c/span\u003e \u003cem\u003esituates these practices within the broader global distribution of AI ethics research\u003c/em\u003e, illustrating their relative marginality compared to dominant Global North frameworks.\u003c/p\u003e \u003cp\u003e \u003cb\u003e3.3. RQ3. What empirical pathways indicate movement from symbolic ethical compliance toward structural redress in AI governance for education?\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe reviewed studies consistently distinguished between \u003cem\u003esymbolic adoption of ethical principles and governance mechanisms capable of producing observable structural effects\u003c/em\u003e. While global AI ethics guidelines\u0026mdash;most notably UNESCO (UNESCO, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2021\u003c/span\u003e)\u0026mdash;were frequently cited, fewer than \u003cb\u003e25%\u003c/b\u003e \u003cem\u003eof studies documented enforceable mechanisms\u003c/em\u003e linking ethical commitments to institutional accountability (Sahebi \u0026amp; Formosa, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eEmpirical synthesis identified three governance pathways associated with more substantive outcomes:\u003c/p\u003e \u003cp\u003e \u003col\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eBias-linked procurement and auditing, where ethical compliance was tied to algorithmic performance evaluations and contractual obligations (Nazer et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Upadhyay et al., \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eInstitutionalized participatory governance, including educator-led review boards and community audit mechanisms, which were associated with improved trust and contextual relevance (Olojede, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eCapacity-building through AI literacy and open infrastructures, enabling sustained local adaptation and oversight (Madlberger, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Ng et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2021b\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2021a\u003c/span\u003e).\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003c/ol\u003e \u003c/p\u003e \u003cp\u003eThese pathways are not presented as normative prescriptions but as empirically observed mechanisms across the reviewed studies. Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e3\u003c/span\u003e captures how such mechanisms remain unevenly distributed across regions, reinforcing the contrast between symbolic compliance and structurally embedded ethics.\u003c/p\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e3.4. Integrative Empirical Synthesis\u003c/h2\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e2\u003c/span\u003e functions as the primary \u003cem\u003ethematic synthesis\u003c/em\u003e of empirical findings, identifying what ethical issues and governance practices recur across AI ethics in education. Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e3\u003c/span\u003e complements this synthesis by providing a \u003cem\u003edistributional analysis\u003c/em\u003e, mapping where these themes are produced, implemented, and governed, and by whom. Together, the tables distinguish between the content of ethical concerns and the structural conditions under which they are institutionalized.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eThematic synthesis of AI Ethics in Education: From Empirical Findings to Policy Implications\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTheme\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRepresentative Studies\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTheoretical Link\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePolicy Implication\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDominance of Global North frameworks in AI ethics\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eKhan et al. (\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2022\u003c/span\u003e); Upadhyay et al. (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2023\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003eColoniality of knowledge\u003c/em\u003e: Euro-American standards universalized as \u0026ldquo;global\u0026rdquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRequire glocal adaptation of guidelines; reserve policy seats for Global South actors\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAlgorithmic exclusion of non-native speakers\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eZajko (\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2021\u003c/span\u003e); Baker \u0026amp; Hawn (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2022\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003eColonial infrastructures\u003c/em\u003e: monolingual datasets reproduce systemic bias\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMandate multilingual datasets and local bias audits in procurement\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLack of participatory governance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUNESCO (\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2021\u003c/span\u003e); Olojede (\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2023\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003eEpistemic disobedience\u003c/em\u003e: absence of local voices shows systemic silencing\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eInstitutionalize community co-design and participatory policy mechanisms\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUbuntu-informed and Indigenous approaches\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAmmah et al. (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2024\u003c/span\u003e); Zembylas (\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2023\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003ePluriversality\u003c/em\u003e: multiple knowledge systems as coequal ethical sources\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eScale up South-led, culturally grounded AI ethics toolkits\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVendor lock-in and dependency\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eArora et al. (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2023\u003c/span\u003e); Sahebi \u0026amp; Formosa (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2024\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003eColoniality of power\u003c/em\u003e: economic dependency embedded in digital infrastructures\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eEnforce open-source alternatives and transparent procurement\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCulturally responsive pedagogy in AI literacy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMadlberger (2017); Vetter (2024)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003eEpistemic justice\u003c/em\u003e: validating local pedagogical traditions\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIntegrate AI ethics into teacher training and curricula in context-sensitive ways\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e3\u003c/span\u003e captures how these governance mechanisms remain unevenly distributed across regions, highlighting structural disparities rather than thematic repetition.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDistributional Mapping of AI Ethics in Education Studies through a Decolonial Lens\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTheme\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRepresentative Studies\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTheoretical Link (Decolonial Lens)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePolicy Implication\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eGeopolitical dominance of Global North\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eKhan et al. (\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) \u0026ndash; majority frameworks Euro-American; Upadhyay, Pasupuleti (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) \u0026ndash; weak Global South representation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003eColoniality of knowledge\u003c/em\u003e: Euro-American epistemologies universalized as \u0026ldquo;global\u0026rdquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDevelop South-led frameworks; allocate research funding to underrepresented regions\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAlgorithmic exclusion \u0026amp; linguistic bias\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eZajko (\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2021\u003c/span\u003e); Baker \u0026amp; Hawn (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) \u0026ndash; ~15% accuracy gap for non-native English speakers\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003eColoniality of language\u003c/em\u003e: English-centric datasets systematically marginalize others\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRequire multilingual datasets; enforce context-aware validation of algorithms\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eTokenistic inclusion without participation\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUNESCO (\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) \u0026ndash; guidelines often adopted symbolically; Sahebi \u0026amp; Formosa (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) \u0026ndash; superficial compliance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003eEpistemic disobedience\u003c/em\u003e absent \u0026rarr; participation remains formal, not substantive\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eInstitutionalize stakeholder quotas; mandate community co-design in EdTech projects\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eEmerging pluriversal counter-models\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAmmah et al. (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) \u0026ndash; Ubuntu AI ethics; Madlberger (2017)\u0026ndash; localized pedagogies\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003ePluriversality\u003c/em\u003e: diverse epistemologies as legitimate ethical foundations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eScale up co-created curricula; fund pluriversal AI literacy and governance toolkits\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cem\u003eThe complete table is presented in the Appendix (\u003c/em\u003eTable\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e3\u003c/span\u003eA\u003cem\u003e).\u003c/em\u003e\u003c/p\u003e \u003cp\u003eThis table presents a distributional analysis of empirical findings across the reviewed studies, mapping where AI ethics frameworks in education are produced, whose epistemic perspectives dominate, and how governance authority is allocated across geopolitical contexts. Unlike the thematic synthesis in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e2\u003c/span\u003e, which consolidates recurrent ethical issues and practices, Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e3\u003c/span\u003e highlights structural asymmetries in authorship, datasets, linguistic orientation, and decision-making power. The table provides empirical evidence of how ethical principles and governance mechanisms are unevenly institutionalized, revealing persistent concentrations of authority in Global North contexts alongside the relative marginality of pluriversal and locally grounded approaches.\u003c/p\u003e \u003cp\u003e \u003cb\u003eResults Summary\u003c/b\u003e \u003c/p\u003e \u003cp\u003eIn summary, the Results indicate that:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eAI ethics in education is predominantly shaped by Global North\u0026ndash;produced frameworks with limited contextual adaptation.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eAlgorithmic bias and governance exclusion are empirically documented and systematically patterned.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003ePluriversal and localized practices constitute observable ethical alternatives rather than isolated exceptions.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eStructural redress is empirically associated with governance mechanisms that embed ethics into participation, procurement, and capacity-building processes.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eInterpretation of these findings, including their theoretical implications and the development of the Framework for \u003cb\u003eFGAIEE\u003c/b\u003e, is presented in the following Discussion section.\u003c/p\u003e \u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eRecent scholarship on AI ethics in education has highlighted the limitations of principle-based and universalist frameworks. Although influential models have established shared vocabularies of fairness, transparency, and accountability (du Boulay, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Holmes et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), empirical evidence shows that these principles often remain weakly connected to institutional governance and pedagogical realities, particularly beyond the Global North. Consistent with the findings of this review, ethical commitments alone rarely translate into enforceable or contextually legitimate practices in educational AI systems. Taken together, these findings reposition AI ethics in education as a governance problem rooted in epistemic power and institutional design, rather than a deficit of ethical principles.\u003c/p\u003e \u003cp\u003eWhile prior work in AI and education has addressed ethical principles and pedagogical competence (Holmes et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Wiese et al., \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), governance-focused reviews reveal that ethical guidelines often lack enforceable mechanisms (Batool et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). These observations align with our findings and underscore the need for frameworks that bridge global principles with local legitimacy \u0026mdash; a gap addressed by the \u003cb\u003eFGAIEE\u003c/b\u003e.\u003c/p\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e4.1. Reframing AI Ethics in Education as an Epistemic Power Structure\u003c/h2\u003e \u003cp\u003eThe findings reported in RQ1 confirm that AI ethics in education operates not merely as a normative or technical domain, but as an \u003cem\u003eepistemic power structure\u003c/em\u003e. The dominance of Global North\u0026ndash;produced ethical frameworks reflects what Quijano (\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2000\u003c/span\u003e) conceptualized as the \u003cem\u003ecoloniality of power\u003c/em\u003e, wherein authority over knowledge production persists beyond formal colonialism (Newman, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). When ethical legitimacy is defined externally\u0026mdash;through Euro-American moral philosophies, Anglophone datasets, and universalist governance templates\u0026mdash;AI ethics risks reproducing epistemic hierarchies under the guise of neutrality.\u003c/p\u003e \u003cp\u003eThis observation aligns with Santos\u0026rsquo; (2023) critique of \u003cem\u003eepistemicide\u003c/em\u003e, in which alternative knowledge systems are rendered invisible by claims of universality. The Results demonstrate that global AI ethics frameworks in education often function as regulatory abstractions, disconnected from lived pedagogical realities. As evidenced by the lack of participatory governance and persistent algorithmic bias, ethics becomes procedural rather than transformative.\u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e4\u003c/span\u003e deepens this interpretation by demonstrating how principle-based ethics frameworks translate into symbolic compliance precisely because they are decoupled from decision-making authority and enforcement mechanisms. The table illustrates that ethical commitments remain largely aspirational unless embedded in enforceable institutional mechanisms. This finding resonates with critiques of \u0026ldquo;ethics washing\u0026rdquo; in AI governance (Bietti, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Papyshev \u0026amp; Chan, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), extending them specifically into educational contexts.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePractical Leverage Points for Ethical AI in Education\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLeverage Point\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIllustrative Examples / Literature\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIntegration with FGAIEE\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(1) \u003cb\u003e50% representation of local communities in AI design\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003eGotong royong\u003c/em\u003e deliberation; Ubuntu ethics; \u003cem\u003evanua\u003c/em\u003e governance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEnsures epistemic inclusion in governance processes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(2) \u003cb\u003eMandatory bias audits \u0026amp; diverse data engagement\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eBias reduction initiatives (Nazer et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2023\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEmbeds accountability and transparency into procurement\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(3) \u003cb\u003ePublic\u0026ndash;private partnerships for low-cost AI in LMICs\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInfrastructure pilots in resource-constrained schools\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLinks equity to resource redistribution\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(4) \u003cb\u003eIntegration of AI literacy into curricula \u0026amp; teacher training\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFour-pillar approach (Ng et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2021a\u003c/span\u003e, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003eb\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePromotes ethical awareness as a systemic outcome\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(5) \u003cb\u003eLocal culture-based participatory design\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDigital storytelling (Davy Tsz Kit et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2022\u003c/span\u003e); ethical co-design (Vetter et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2024\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEmbeds cultural values as ethical foundations\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(6) \u003cb\u003eSouth-led AI ethics research funding\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIncrease in Global South publications and grants\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eStrengthens epistemic sovereignty in knowledge production\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(7) \u003cb\u003eGlocalized governance model\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUNESCO\u0026rsquo;s ROAM-X pillars adapted locally (Papakostas, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2025\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eOperationalizes global\u0026ndash;local co-ownership of standards\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(8) \u003cb\u003eInclusion- and disability-friendly design (frugal AI)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSMS-based AI, local-language platforms, inclusive UI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEnsures accessibility as a marker of ethical AI\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eFrom Universalism to Pluriversality\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe empirical patterns identified in RQ2 provide strong support for \u003cem\u003epluriversality\u003c/em\u003e as a necessary theoretical shift in AI ethics. Drawing on Mignolo (\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), pluriversality rejects the assumption that ethical universals can be derived from a single epistemic center. Instead, it affirms the coexistence of multiple ethical rationalities. Where participatory governance concerns \u003cem\u003ewho decides\u003c/em\u003e, epistemic pluriversality addresses \u003cem\u003ewhose knowledge defines what counts as ethical in the first place\u003c/em\u003e.\u003c/p\u003e \u003cp\u003eThese findings align with emerging Global South\u0026ndash;centered scholarship that foregrounds communal and relational ethics in AI governance. Studies in African higher education contexts emphasize Ubuntu as a normative foundation for AI ethics, positioning ethical responsibility as collective stewardship rather than individual compliance (Corrigan et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Yilma, \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Such perspectives empirically reinforce the argument that ethical legitimacy in education is culturally embedded and epistemically plural.\u003c/p\u003e \u003cp\u003ePractices grounded in Ubuntu ethics, \u003cem\u003egotong royong\u003c/em\u003e, and Indigenous data sovereignty exemplify what Zembylas (\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) describes as \u003cem\u003erelational ethics\u003c/em\u003e\u0026mdash;ethical reasoning embedded in social responsibility, reciprocity, and collective care. These findings challenge dominant liberal-individualist ethics frameworks that prioritize autonomy and consent detached from community context.\u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e functions as an analytical pivot, translating the empirical patterns identified in RQ2 into a coherent model of how local ethical traditions interact with global AI infrastructures. Rather than positioning local ethics as subordinate or supplementary, the figure conceptualizes them as \u003cem\u003eco-constitutive layers\u003c/em\u003e of ethical governance. This reframing moves beyond cultural \u0026ldquo;adaptation\u0026rdquo; toward epistemic parity.\u003c/p\u003e \u003cp\u003eThe dialog between Results and theory here confirms that pluriversal ethics are not merely philosophical ideals but \u003cem\u003eempirically observable governance practices\u003c/em\u003e. As such, they provide a credible foundation for rethinking AI ethics in education at scale.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThis figure synthesizes empirical patterns identified across RQ1\u0026ndash;RQ3, illustrating the transition from principle-based and symbolic AI ethics toward structurally embedded governance mechanisms. The model highlights how participatory governance, epistemic inclusion, and institutional accountability function as empirical pathways enabling ethical principles to translate into legitimate and enforceable practices in educational AI systems.\u003c/p\u003e \u003cp\u003e \u003cb\u003eOperationalizing Ethics: From Principles to Governance Mechanisms\u003c/b\u003e \u003c/p\u003e \u003cp\u003eOne of the most significant contributions of this study lies in clarifying the distinction between \u003cem\u003eethical principles\u003c/em\u003e and \u003cem\u003eethical mechanisms\u003c/em\u003e. The Results reveal that while ethical principles are widely cited, they rarely translate into enforceable institutional action. This gap echoes Heeks\u0026rsquo; (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) argument that digital development initiatives often fail due to an \u0026ldquo;implementation void\u0026rdquo; between intention and practice.\u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e5\u003c/span\u003e plays a critical role in bridging this gap by identifying empirically grounded governance instruments\u0026mdash;such as bias-linked procurement, participatory audits, and educator-led oversight\u0026mdash;that operationalize ethical commitments. Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e5\u003c/span\u003e does not introduce new empirical findings but synthesizes observed disparities into a comparative governance heuristic that clarifies pathways for structural enforceability. These mechanisms resonate with Ostrom\u0026rsquo;s (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e1990\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2010\u003c/span\u003e) and Carlisle \u0026amp; Gruby (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) theory of \u003cem\u003epolycentric governance\u003c/em\u003e, which emphasizes distributed authority and collective rule-making over centralized control.\u003c/p\u003e \u003cp\u003eBy situating ethics within procurement contracts, audit procedures, and institutional accountability structures, these mechanisms shift AI ethics from moral persuasion to \u003cem\u003estructural enforceability\u003c/em\u003e. This distinction is essential for education systems, where AI tools directly influence assessment, access, and learner trajectories.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eHeatmap: Regional Disparities in AI Ethics Implementation\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRegion\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInfrastructure\u003c/p\u003e \u003cp\u003eReadiness\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePolicy\u003c/p\u003e \u003cp\u003eStrength\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCultural\u003c/p\u003e \u003cp\u003eAdaptability\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eStakeholder\u003c/p\u003e \u003cp\u003eInclusion\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSub-Saharan Africa\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLow (●○○)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMedium (●●○)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHigh (●●●)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eLow (●○○)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSouth Asia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMedium (●●○)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLow (●○○)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMedium (●●○)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eMedium (●●○)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEU/North America\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHigh (●●●)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHigh (●●●)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLow (●○○)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHigh (●●●)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLatin America\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMedium (●●○)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMedium (●●○)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHigh (●●●)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eLow (●○○)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003e\u003cem\u003eExplanation\u003c/em\u003e:\u003c/td\u003e\u003c/tr\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003e\u0026bull; ●●● = High/Strong, ●●○ = Medium, ●○○ = Low\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003e \u003cem\u003eSource: Synthesis of\u003c/em\u003e Yadav et al. (\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), Wakunuma \u0026amp; Eke (\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2024\u003c/span\u003e\u003cem\u003e), and\u003c/em\u003e Shams et al. (\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eInsight\u003c/b\u003e:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eGlobal North-South Pattern: Strong infrastructure and policy readiness in the North, better cultural adaptation in the South.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eCritical Gap: Low stakeholder inclusion in the Global South despite high cultural adaptation.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eSimilar critiques have been articulated in recent work on decolonising educational technology, which argues that ethical AI in education requires structural reconfiguration of governance, not merely ethical adaptation of existing technologies (Koole et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e4.2. The FGAIEE Framework: Toward Glocalized and Enforceable AI Ethics in Education\u003c/h2\u003e \u003cp\u003eThe empirical findings of this review indicate that prevailing AI ethics frameworks in education struggle to translate universal ethical principles into contextually legitimate and enforceable governance practices. As demonstrated in the Results (RQ1\u0026ndash;RQ3), ethical failures are not incidental but structurally produced through epistemic centralization, limited participation, and weak accountability mechanisms. In response, this study advances the \u003cb\u003eFGAIEE\u003c/b\u003e as an integrative theoretical synthesis that bridges empirical evidence with decolonial and governance theory.\u003c/p\u003e \u003cp\u003eFGAIEE is grounded in the proposition that \u003cem\u003eethical legitimacy in AI-mediated education emerges through negotiated alignment between global infrastructures and local epistemologies\u003c/em\u003e, rather than through the uncritical adoption of universal norms. This positioning extends decolonial critiques of digital colonialism (Quijano, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2000\u003c/span\u003e; Santos, 2014; Mignolo, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) by translating them into an operational framework applicable to educational governance.\u003c/p\u003e \u003cp\u003e \u003cb\u003eEpistemic Pluriversality as the Ethical Foundation\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe first pillar of FGAIEE is \u003cem\u003eepistemic pluriversality\u003c/em\u003e, which recognizes multiple knowledge systems as coequal sources of ethical authority. The Results demonstrate that Global North\u0026ndash;centric ethics frameworks frequently marginalize local pedagogical values, linguistic norms, and culturally embedded assessment practices. FGAIEE responds by repositioning local ethical traditions\u0026mdash;such as Ubuntu ethics, \u003cem\u003egotong royong\u003c/em\u003e, Indigenous data sovereignty, and religious epistemologies\u0026mdash;not as contextual adaptations, but as foundational inputs into ethical governance.\u003c/p\u003e \u003cp\u003eAs illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, epistemic pluriversality operates as the base layer of the framework, shaping how ethical questions are defined, deliberated, and resolved. This approach aligns with relational ethics perspectives (Zembylas, \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) and directly addresses the epistemic exclusion observed across the reviewed studies. Rather than seeking ethical uniformity, FGAIEE enables \u003cem\u003eethical coordination without epistemic domination\u003c/em\u003e.\u003c/p\u003e \u003cp\u003e \u003cb\u003eParticipatory and Polycentric Governance\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe second pillar operationalizes ethics through \u003cem\u003eparticipatory and polycentric governance structures\u003c/em\u003e. Consistent with the empirical patterns identified in RQ1 and RQ2, the absence of meaningful stakeholder participation correlates with ethical fragility and low institutional trust. Drawing on Ostrom\u0026rsquo;s (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e1990\u003c/span\u003e) theory of polycentric governance, FGAIEE distributes ethical authority across multiple actors and levels, including educators, learners, communities, institutions, and policymakers.\u003c/p\u003e \u003cp\u003eThe centrality of educator participation identified in this study is consistent with recent empirical research showing that educators perceive AI ethics as a complex, context-dependent challenge requiring institutional support rather than abstract guidance (Kamali, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). This reinforces the framework\u0026rsquo;s emphasis on participatory and polycentric governance as a prerequisite for ethical sustainability in educational AI systems.\u003c/p\u003e \u003cp\u003eThis governance logic is reflected in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e4\u003c/span\u003e, which synthesizes empirically documented mechanisms such as educator-led ethics boards, community review processes, and participatory audits. These mechanisms demonstrate that ethical legitimacy increases when those affected by AI systems are directly involved in their oversight. Within FGAIEE, participation is not consultative but \u003cem\u003edecision-oriented\u003c/em\u003e, ensuring that ethical governance remains context-sensitive and contestable.\u003c/p\u003e \u003cp\u003e \u003cb\u003eStructural Enforceability of Ethical Commitments\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe third pillar addresses the most persistent weakness of existing AI ethics frameworks: the lack of enforceability. The Results reveal that ethical principles frequently remain symbolic unless embedded within institutional mechanisms. FGAIEE therefore conceptualizes ethics as a \u003cem\u003estructural property of governance\u003c/em\u003e, not a voluntary moral stance.\u003c/p\u003e \u003cp\u003eAs consolidated in Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e5\u003c/span\u003e, enforceability is operationalized through bias-linked procurement requirements, mandatory algorithmic audits, accountability clauses, and sanctions for non-compliance. These instruments transform ethics into an observable and evaluable practice, aligning with critiques of \u0026ldquo;ethics washing\u0026rdquo; in AI governance (Metcalf et al., 2019; Bietti, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). In educational contexts\u0026mdash;where AI systems directly affect assessment, access, and learner trajectories\u0026mdash;structural enforceability is essential for ethical credibility.\u003c/p\u003e \u003cp\u003e \u003cb\u003eMulti-Scalar Architecture and Integration\u003c/b\u003e \u003c/p\u003e \u003cp\u003eFGAIEE operates across interconnected global, institutional, and local scales. Figure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e visualizes this multi-scalar architecture, demonstrating how global standards and infrastructures interact dynamically with institutional policies and local ethical deliberation. Importantly, no single scale is privileged; ethical alignment emerges through continuous negotiation rather than top-down imposition. By explicitly linking regional disparities to governance mechanisms and outcomes, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e advances existing AI ethics models by operationalizing decolonial theory into a multi-scalar governance pathway.\u003c/p\u003e \u003cp\u003eThis architecture explains why purely global or purely local ethics frameworks are insufficient. FGAIEE instead offers a \u003cem\u003eprocedural model\u003c/em\u003e that enables contextual articulation of ethics while maintaining global interoperability\u0026mdash;an approach particularly suited to education systems characterized by cultural diversity and infrastructural inequality.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe figure illustrates FGAIEE as an integrative framework connecting epistemic justice, participatory governance, and structural accountability across global\u0026ndash;local educational contexts. Derived from empirical synthesis, the framework demonstrates how ethical principles in AI-enabled education become institutionally legitimate and operational through context-sensitive governance mechanisms.\u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e depicts a decolonial governance model that operationalizes glocalized ethics by translating regional disparities (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e5\u003c/span\u003e) into strategies for structural redress. At the input stage, AI ethics in education is shaped by tensions between Global North frameworks, local epistemologies, and contextual constraints. Through PICo\u0026ndash;SWOT synthesis, moderated by decolonial lenses\u0026mdash;\u003cem\u003ecoloniality of knowledge, epistemic disobedience, pluriversality\u003c/em\u003e\u0026mdash;the model identifies asymmetries and openings for reform. These are activated by mediators such as local policy seats, participatory co-design, and culturally responsive teacher training, which ensure epistemic inclusion. The resulting outputs\u0026mdash;bias audits, South-led open-source toolkits, and glocalized ethical standards\u0026mdash;produce outcomes of structural redress, shared authorship, and democratized educational futures. By tracing this causal pathway, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e advances beyond principle-based universalism (Floridi et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; UNESCO, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) and reframes AI ethics in education as a multi-scalar political project grounded in redistribution and epistemic justice.\u003c/p\u003e \u003cp\u003e \u003cb\u003ePositioning FGAIEE as a Theoretical Contribution\u003c/b\u003e \u003c/p\u003e \u003cp\u003eSynthesizing empirical findings and theory, FGAIEE advances AI ethics in education in three key ways. First, it reframes ethics as an epistemic and governance challenge rather than a checklist of principles. Second, it integrates decolonial theory with institutional design, translating critique into actionable governance. Third, it provides a transferable framework based on \u003cb\u003eprocess adaptability\u003c/b\u003e, not value standardization.\u003c/p\u003e \u003cp\u003eWithin the structure of this article, FGAIEE functions as the \u003cb\u003econceptual resolution of RQ3\u003c/b\u003e, offering a coherent explanation of how AI ethics in education can move from symbolic compliance toward structural redress. Rather than proposing another universal model, the framework redefines ethical AI as a shared, participatory, and enforceable project shaped across epistemic and geopolitical boundaries.\u003c/p\u003e \u003cp\u003e \u003cb\u003eLimitations and Future Directions\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThis review, while offering a broad synthesis of AI ethics in education, has several limitations. First, its temporal scope (2015\u0026ndash;2025/fraction) may not fully capture rapid advances in generative AI since 2022. Future studies should include real-time updates via preprints and continuous mapping. Second, language and geographic biases persist most literature stems from English-speaking, Global North institutions. Expanding language diversity and supporting open-access publishing in the Global South will help decentralize knowledge production. Third, the analysis is based on secondary data, limiting insights into real-world implementation. Future research should include context-rich, empirical case studies, e.g., Ubuntu-based AI curricula in Africa or ethics rooted in \u003cem\u003egotong royong\u003c/em\u003e in Southeast Asia, to better understand culturally grounded practices. These steps are vital for advancing equitable and inclusive AI in education.\u003c/p\u003e \u003c/div\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eThis systematic review demonstrates that AI ethics in education cannot be understood as a neutral or purely technical domain. Rather, it is a contested epistemic field shaped by asymmetries of power, knowledge production, and governance authority. The evidence synthesized across 84 studies reveals that dominant AI ethics frameworks\u0026mdash;largely authored in the Global North\u0026mdash;continue to reproduce the coloniality of knowledge by privileging Anglophone datasets, Euro-American moral assumptions, and universalist policy templates. In educational contexts, these dynamics translate into algorithmic exclusion, symbolic inclusion, and ethical compliance without structural transformation.\u003c/p\u003e \u003cp\u003eAt the same time, this review shows that alternative ethical futures are not hypothetical. Across diverse regions, pluriversal practices grounded in Ubuntu ethics, Indigenous data sovereignty, \u003cem\u003egotong royong\u003c/em\u003e governance, and culturally responsive pedagogies demonstrate that ethical AI in education can be co-produced, contextually legitimate, and socially transformative. These initiatives enact what decolonial scholars describe as \u003cem\u003eepistemic disobedience\u003c/em\u003e: a refusal to accept imported ethical templates as universal, and an assertion of local knowledge systems as legitimate foundations for governance.\u003c/p\u003e \u003cp\u003eThe central theoretical contribution of this study is the \u003cb\u003eFramework for Glocalized AI Ethics in Education (FGAIEE)\u003c/b\u003e. Unlike principle-based universalist models, FGAIEE operationalizes ethics as a process of structural redress. It integrates epistemic pluriversality, participatory governance, and enforceable policy mechanisms into a coherent multi-scalar model that links global standards with local legitimacy. By embedding ethical commitments into procurement processes, institutional audits, curriculum design, and research funding structures, FGAIEE shifts AI ethics in education from aspirational rhetoric to actionable governance.\u003c/p\u003e \u003cp\u003eMethodologically, the study advances a novel PICo\u0026ndash;Thematic\u0026ndash;SWOT synthesis that demonstrates how critical theory can be translated into policy-relevant insights without sacrificing analytical depth. This approach offers a replicable model for future reviews of ethical AI across sectors, particularly in contexts where global frameworks intersect with local realities. Rather than seeking convergence around abstract principles, the analysis shows that \u003cem\u003edivergence, plurality, and contextual specificity\u003c/em\u003e are the true indicators of ethical resilience.\u003c/p\u003e \u003cp\u003ePractically, the findings identify concrete leverage points\u0026mdash;mandatory bias audits, participatory design mandates, South-led research funding, culturally embedded AI literacy, and frugal technological infrastructures\u0026mdash;that can democratize AI governance in education. These measures are not supplementary but essential: without redistributing epistemic authority and institutional power, ethical AI risks becoming an alibi for continued technological expansion rather than a safeguard for justice.\u003c/p\u003e \u003cp\u003eIn conclusion, this study offers both critique and construction. It exposes how current AI ethics regimes in education reproduce digital colonialism, while simultaneously advancing a globally relevant decolonial framework for transformation. By reframing ethics as shared authorship across epistemologies and scales, the Framework for Glocalized AI Ethics in Education positions ethical AI not as a universal checklist, but as an ongoing political project\u0026mdash;one that will determine whose knowledge, values, and futures are encoded into the educational systems of the algorithmic age.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eFunding Statement\u003c/h2\u003e \u003cp\u003eThis research received no external funding.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eDwi Mariyono (DM) conceptualized the study, led the research design, conducted the systematic review and data synthesis, developed the Framework for Glocalized AI Ethics in Education (FGAIEE), and drafted the original manuscript. DM also served as the corresponding author and coordinated all stages of manuscript development and revision.Muhammad Yunus (MY) contributed to data screening, methodological refinement, and critical review of the empirical findings. MY provided substantive intellectual input to the analysis and interpretation of results and reviewed the manuscript for conceptual clarity and academic rigor.Akmal Nur Alif Hidayatullah (ANAH) supported data extraction and organization, assisted in thematic mapping and verification of findings, and contributed to the refinement of figures, tables, and manuscript formatting. ANAH also participated in manuscript revision and proofreading.All authors reviewed and approved the final manuscript and agreed to be accountable for all aspects of the work.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eThe authors would like to acknowledge the contributions of the global scholarly community whose work formed the empirical foundation of this review. We are particularly indebted to researchers, educators, and policy scholars whose openly accessible publications enabled a comprehensive and inclusive synthesis across diverse geopolitical and epistemic contexts.We also acknowledge the constructive role of international peer-review standards and open academic infrastructures that support transparent, cumulative, and globally accessible knowledge production in the field of artificial intelligence in education.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eAll data underlying this study are derived from publicly accessible and peer-reviewed sources. The reviewed materials consist exclusively of journal articles, reports, and policy documents indexed in internationally recognized academic databases and repositories, including SpringerLink, Scopus, Web of Science, ERIC, and publisher websites of international organizations.No proprietary datasets were used. All sources analyzed in this review are cited in the reference list and can be accessed online without restriction, ensuring full transparency and reproducibility of the review process.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAbdalla, S., Abdeh Kolahchi, A., Ablain, M., Adusumilli, S., Aich Bhowmick, S., Alou-Font, E., Amarouche, L., Andersen, O. B., Antich, H., Aouf, L., Arbic, B., Armitage, T., Arnault, S., Artana, C., Aulicino, G., Ayoub, N., Badulin, S., Baker, S., Banks, C., \u0026amp; Zlotnicki, V. (2021). Altimetry for the future: Building on 25 years of progress. \u003cem\u003eAdvances in Space Research\u003c/em\u003e, \u003cem\u003e68\u003c/em\u003e(2), 319\u0026ndash;363. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.asr.2021.01.022\u003c/span\u003e\u003cspan address=\"10.1016/j.asr.2021.01.022\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmmah, L. N. A., L\u0026uuml;tge, C., Kriebitz, A., \u0026amp; Ramkissoon, L. (2024). AI4people\u0026thinsp;\u0026ndash;\u0026thinsp;an ethical framework for a good AI society: the Ghana (Ga) perspective. \u003cem\u003eJournal of Information Communication and Ethics in Society\u003c/em\u003e, \u003cem\u003e22\u003c/em\u003e(4), 453\u0026ndash;465. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1108/JICES-06-2024-0072\u003c/span\u003e\u003cspan address=\"10.1108/JICES-06-2024-0072\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArora, A., Barrett, M., Lee, E., Oborn, E., \u0026amp; Prince, K. (2023). Risk and the future of AI: Algorithmic bias, data colonialism, and marginalization. \u003cem\u003eInformation and Organization\u003c/em\u003e, \u003cem\u003e33\u003c/em\u003e(3), 100478. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.infoandorg.2023.100478\u003c/span\u003e\u003cspan address=\"10.1016/j.infoandorg.2023.100478\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaker, R. S., \u0026amp; Hawn, A. (2022). Algorithmic Bias in Education. \u003cem\u003eInternational Journal of Artificial Intelligence in Education\u003c/em\u003e, \u003cem\u003e32\u003c/em\u003e(4), 1052\u0026ndash;1092. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s40593-021-00285-9\u003c/span\u003e\u003cspan address=\"10.1007/s40593-021-00285-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBatool, A., Zowghi, D., \u0026amp; Bano, M. (2025). AI governance: a systematic literature review. \u003cem\u003eAI and Ethics\u003c/em\u003e, \u003cem\u003e5\u003c/em\u003e(3), 3265\u0026ndash;3279. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s43681-024-00653-w\u003c/span\u003e\u003cspan address=\"10.1007/s43681-024-00653-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBietti, E. (2020). From ethics washing to ethics bashing. \u003cem\u003eProceedings of the 2020 Conference on Fairness, Accountability, and Transparency\u003c/em\u003e, 210\u0026ndash;219. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1145/3351095.3372860\u003c/span\u003e\u003cspan address=\"10.1145/3351095.3372860\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBirhane, A. (2021). Algorithmic injustice: a relational ethics approach. \u003cem\u003ePatterns\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(2), 100205. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.patter.2021.100205\u003c/span\u003e\u003cspan address=\"10.1016/j.patter.2021.100205\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBraun, V., \u0026amp; Clarke, V. (2006). Qualitative Research in Psychology Using thematic analysis in psychology Using thematic analysis in psychology. \u003cem\u003eQualitative Research in Psychology\u003c/em\u003e, \u003cem\u003e3\u003c/em\u003e(2), 77\u0026ndash;101. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.tandfonline.com/action/journalInformation?journalCode=uqrp20%5Cnhttp://www.tandfonline.com/action/journalInformation?journalCode=uqrp20\u003c/span\u003e\u003cspan address=\"http://www.tandfonline.com/action/journalInformation?journalCode=uqrp20%5Cnhttp://www.tandfonline.com/action/journalInformation?journalCode=uqrp20\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBraun, V., \u0026amp; Clarke, V. (2021). Conceptual and Design Thinking for Thematic Analysis. \u003cem\u003eQualitative Psychology\u003c/em\u003e, \u003cem\u003e9\u003c/em\u003e(1), 3\u0026ndash;26. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1037/qup0000196\u003c/span\u003e\u003cspan address=\"10.1037/qup0000196\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarlisle, K., \u0026amp; Gruby, R. L. (2019). Polycentric Systems of Governance: A Theoretical Model for the Commons. \u003cem\u003ePolicy Studies Journal\u003c/em\u003e, \u003cem\u003e47\u003c/em\u003e(4), 927\u0026ndash;952. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/psj.12212\u003c/span\u003e\u003cspan address=\"10.1111/psj.12212\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCorrigan, C. C., Asakipaam, S. A., Kponyo, J. J., \u0026amp; Luetge, C. (Eds.). (2023). \u003cem\u003eAI Ethics in Higher Education: Insights from Africa and Beyond\u003c/em\u003e. Springer International Publishing. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-3-031-23035-6\u003c/span\u003e\u003cspan address=\"10.1007/978-3-031-23035-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCouldry, N., \u0026amp; Mejias, U. A. (2019). Data Colonialism: Rethinking Big Data\u0026rsquo;s Relation to the Contemporary Subject. \u003cem\u003eTelevision \u0026amp; New Media\u003c/em\u003e, \u003cem\u003e20\u003c/em\u003e(4), 336\u0026ndash;349. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/1527476418796632\u003c/span\u003e\u003cspan address=\"10.1177/1527476418796632\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDavy Tsz Kit, N. G., Luo, W., Chan, H. M. Y., \u0026amp; Chu, S. K. W. (2022). Using digital story writing as a pedagogy to develop AI literacy among primary students. \u003cem\u003eComputers and Education: Artificial Intelligence\u003c/em\u003e, \u003cem\u003e3\u003c/em\u003e, 100054. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.caeai.2022.100054\u003c/span\u003e\u003cspan address=\"10.1016/j.caeai.2022.100054\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ede Sousa Santos, B. (2023). Epistemologies of the south: Justice against epistemicide. In \u003cem\u003eFoundational Concepts of Decolonial and Southern Epistemologies\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/0094306115609925nn\u003c/span\u003e\u003cspan address=\"10.1177/0094306115609925nn\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003edu Boulay, B. (2023). Artificial Intelligence in Education and Ethics. In \u003cem\u003eHandbook of Open, Distance and Digital Education\u003c/em\u003e (pp. 93\u0026ndash;108). Springer Nature Singapore. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-981-19-2080-6_6\u003c/span\u003e\u003cspan address=\"10.1007/978-981-19-2080-6_6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFloridi, L., Cowls, J., Beltrametti, M., Chatila, R., Chazerand, P., Dignum, V., Luetge, C., Madelin, R., Pagallo, U., Rossi, F., Schafer, B., Valcke, P., \u0026amp; Vayena, E. (2018). AI4People\u0026mdash;An Ethical Framework for a Good AI Society: Opportunities, Risks, Principles, and Recommendations. \u003cem\u003eMinds and Machines\u003c/em\u003e, \u003cem\u003e28\u003c/em\u003e(4), 689\u0026ndash;707. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s11023-018-9482-5\u003c/span\u003e\u003cspan address=\"10.1007/s11023-018-9482-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHeeks, R. (2022). Digital inequality beyond the digital divide: conceptualizing adverse digital incorporation in the global South. \u003cem\u003eInformation Technology for Development\u003c/em\u003e, \u003cem\u003e28\u003c/em\u003e(4), 688\u0026ndash;704. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/02681102.2022.2068492\u003c/span\u003e\u003cspan address=\"10.1080/02681102.2022.2068492\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHelms, M. M., \u0026amp; Nixon, J. (2010). Exploring SWOT analysis \u0026ndash; where are we now? A review of academic research from the last decade. \u003cem\u003eJournal of Strategy and Management\u003c/em\u003e, \u003cem\u003e3\u003c/em\u003e(3), 215\u0026ndash;251. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1108/17554251011064837\u003c/span\u003e\u003cspan address=\"10.1108/17554251011064837\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHolmes, W., Porayska-Pomsta, K., Holstein, K., Sutherland, E., Baker, T., Shum, S. B., Santos, O. C., Rodrigo, M. T., Cukurova, M., Bittencourt, I. I., \u0026amp; Koedinger, K. R. (2022). Ethics of AI in Education: Towards a Community-Wide Framework. \u003cem\u003eInternational Journal of Artificial Intelligence in Education\u003c/em\u003e, \u003cem\u003e32\u003c/em\u003e(3), 504\u0026ndash;526. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s40593-021-00239-1\u003c/span\u003e\u003cspan address=\"10.1007/s40593-021-00239-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJobin, A., Ienca, M., \u0026amp; Vayena, E. (2019). The global landscape of AI ethics guidelines. \u003cem\u003eNature Machine Intelligence\u003c/em\u003e, \u003cem\u003e1\u003c/em\u003e(9), 389\u0026ndash;399. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s42256-019-0088-2\u003c/span\u003e\u003cspan address=\"10.1038/s42256-019-0088-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKamali, J. (2024). AI ethics as a complex and multifaceted challenge: decoding educators\u0026rsquo; AI ethics alignment through the lens of activity theory. \u003cem\u003eInternational Journal of Educational Technology in Higher Education\u003c/em\u003e, \u003cem\u003e21\u003c/em\u003e(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s41239-024-00496-9\u003c/span\u003e\u003cspan address=\"10.1186/s41239-024-00496-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKhan, A. A., Badshah, S., Liang, P., Waseem, M., Khan, B., Ahmad, A., Fahmideh, M., Niazi, M., \u0026amp; Akbar, M. A. (2022). Ethics of AI: A Systematic Literature Review of Principles and Challenges. \u003cem\u003eThe International Conference on Evaluation and Assessment in Software Engineering 2022\u003c/em\u003e, 383\u0026ndash;392. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1145/3530019.3531329\u003c/span\u003e\u003cspan address=\"10.1145/3530019.3531329\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKoole, M., Smith, M., Traxler, J., Adam, T., \u0026amp; Footring, S. (2024). Decolonising Educational Technology. \u003cem\u003eEducation Sciences\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(10), 1070. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/educsci14101070\u003c/span\u003e\u003cspan address=\"10.3390/educsci14101070\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKwet, M. (2019). Digital colonialism: US empire and the new imperialism in the Global South. \u003cem\u003eRace \u0026amp; Class\u003c/em\u003e, \u003cem\u003e60\u003c/em\u003e(4), 3\u0026ndash;26. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/0306396818823172\u003c/span\u003e\u003cspan address=\"10.1177/0306396818823172\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMadlberger, L. (2017). Gotong royong in NLP research a mobile tool for collaborative text annotation in Indonesia. In \u003cem\u003eProceedings of the 2016 International Conference on Asian Language Processing Ialp 2016\u003c/em\u003e (pp. 99\u0026ndash;102). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/IALP.2016.7875944\u003c/span\u003e\u003cspan address=\"10.1109/IALP.2016.7875944\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcHugh, M. L. (2012). Interrater reliability: the kappa statistic. \u003cem\u003eBiochemia Medica\u003c/em\u003e, 276\u0026ndash;282. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.11613/BM.2012.031\u003c/span\u003e\u003cspan address=\"10.11613/BM.2012.031\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMehrabi, N., Morstatter, F., Saxena, N., Lerman, K., \u0026amp; Galstyan, A. (2022). A Survey on Bias and Fairness in Machine Learning. \u003cem\u003eACM Computing Surveys\u003c/em\u003e, \u003cem\u003e54\u003c/em\u003e(6), 1\u0026ndash;35. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1145/3457607\u003c/span\u003e\u003cspan address=\"10.1145/3457607\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMignolo, W. D. (2020). The Darker Side of Western Modernity. \u003cem\u003eThe Darker Side of Western Modernity\u003c/em\u003e. Duke University Press. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/j.ctv125jqbw\u003c/span\u003e\u003cspan address=\"10.2307/j.ctv125jqbw\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMohamed, S., Png, M. T., \u0026amp; Isaac, W. (2020). Decolonial AI: Decolonial Theory as Sociotechnical Foresight in Artificial Intelligence. \u003cem\u003ePhilosophy \u0026amp; Technology\u003c/em\u003e, \u003cem\u003e33\u003c/em\u003e(4), 659\u0026ndash;684. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s13347-020-00405-8\u003c/span\u003e\u003cspan address=\"10.1007/s13347-020-00405-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNazer, L. H., Zatarah, R., Waldrip, S., Ke, J. X. C., Moukheiber, M., Khanna, A. K., Hicklen, R. S., Moukheiber, L., Moukheiber, D., Ma, H., \u0026amp; Mathur, P. (2023). Bias in artificial intelligence algorithms and recommendations for mitigation. \u003cem\u003ePLOS Digital Health\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(6), e0000278. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1371/journal.pdig.0000278\u003c/span\u003e\u003cspan address=\"10.1371/journal.pdig.0000278\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNewman, S. J. (2024). An\u0026iacute;bal Quijano\u0026rsquo;s Critical Sociology: From Dependency Theory to Coloniality. \u003cem\u003eSociology of Race and Ethnicity\u003c/em\u003e, \u003cem\u003e10\u003c/em\u003e(4), 566\u0026ndash;571. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/23326492241268532\u003c/span\u003e\u003cspan address=\"10.1177/23326492241268532\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNg, D. T. K., Leung, J. K. L., Chu, K. W. S., \u0026amp; Qiao, M. S. (2021a). AI Literacy: Definition, Teaching, Evaluation and Ethical Issues. \u003cem\u003eProceedings of the Association for Information Science and Technology\u003c/em\u003e, \u003cem\u003e58\u003c/em\u003e(1), 504\u0026ndash;509. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/pra2.487\u003c/span\u003e\u003cspan address=\"10.1002/pra2.487\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNg, D. T. K., Leung, J. K. L., Chu, S. K. W., \u0026amp; Qiao, M. S. (2021b). Conceptualizing AI literacy: An exploratory review. \u003cem\u003eComputers and Education: Artificial Intelligence\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e, 100041. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.caeai.2021.100041\u003c/span\u003e\u003cspan address=\"10.1016/j.caeai.2021.100041\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNurhafida, Y. H., Muchtarom, M., \u0026amp; Gunawati, D. (2020). Strategi FKPM (Forum Kemitraan Polri dan Masyarakat) dalam Meningkatkan Kesadaran Hukum Masyarakat (Studi di Kelurahan Kadipiro Kecamatan Banjarsari Kota Surakarta). \u003cem\u003eJurnal PPKn\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e(1), 1\u0026ndash;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/https://doi.org/10.2019/jppkn.v8i1.73\u003c/span\u003e\u003cspan address=\"10.2019/jppkn.v8i1.73\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eO\u0026rsquo;cathain, A., Murphy, E., \u0026amp; Nicholl, J. (2008). The Quality of Mixed Methods Studies in Health Services Research. \u003cem\u003eJournal of Health Services Research \u0026amp; Policy\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(2), 92\u0026ndash;98. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1258/jhsrp.2007.007074\u003c/span\u003e\u003cspan address=\"10.1258/jhsrp.2007.007074\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOECD (2023). \u003cem\u003eOECD Digital Education Outlook 2023\u003c/em\u003e. OECD. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1787/589b283f-en\u003c/span\u003e\u003cspan address=\"10.1787/589b283f-en\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOlojede, H. T. (2023). Towards African Artificial Intelligence Ethical Principles. \u003cem\u003e2023 First International Conference on the Advancements of Artificial Intelligence in African Context (AAIAC)\u003c/em\u003e, 1\u0026ndash;6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/AAIAC60008.2023.10465379\u003c/span\u003e\u003cspan address=\"10.1109/AAIAC60008.2023.10465379\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOstrom, E. (1990). Governing the commons: the evolution of institutions for collective action. In \u003cem\u003eGoverning the commons: the evolution of institutions for collective action\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/3146384\u003c/span\u003e\u003cspan address=\"10.2307/3146384\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOstrom, E. (2010). Beyond Markets and States: Polycentric Governance of Complex Economic Systems. \u003cem\u003eAmerican Economic Review\u003c/em\u003e, \u003cem\u003e100\u003c/em\u003e(3), 641\u0026ndash;672. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1257/aer.100.3.641\u003c/span\u003e\u003cspan address=\"10.1257/aer.100.3.641\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePage, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., Glanville, J., Grimshaw, J. M., Hr\u0026oacute;bjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-Wilson, E., McDonald, S., \u0026amp; Moher, D. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. \u003cem\u003eInternational Journal of Surgery\u003c/em\u003e, \u003cem\u003e88\u003c/em\u003e, 105906. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ijsu.2021.105906\u003c/span\u003e\u003cspan address=\"10.1016/j.ijsu.2021.105906\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePapakostas, C. (2025). Artificial Intelligence in Religious Education: Ethical, Pedagogical, and Theological Perspectives. \u003cem\u003eReligions\u003c/em\u003e, \u003cem\u003e16\u003c/em\u003e(5), 563. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/rel16050563\u003c/span\u003e\u003cspan address=\"10.3390/rel16050563\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePapyshev, G., \u0026amp; Chan, K. J. D. (2025). Fugazi regulation for AI: strategic tolerance for ethics washing. \u003cem\u003eAI \u0026amp; SOCIETY\u003c/em\u003e, \u003cem\u003e40\u003c/em\u003e(5), 3475\u0026ndash;3486. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00146-024-02084-x\u003c/span\u003e\u003cspan address=\"10.1007/s00146-024-02084-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePasupuleti, M. K. (2024). Ethical AI Governance: A Global Blueprint. In \u003cem\u003eAI Governance: Innovating with Integrity\u003c/em\u003e (pp. 60\u0026ndash;77). National Education Services. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.62311/nesx/97832\u003c/span\u003e\u003cspan address=\"10.62311/nesx/97832\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQuijano, A. (2000). Coloniality of Power and Eurocentrism in Latin America. \u003cem\u003eInternational Sociology\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(2), 215\u0026ndash;232. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/0268580900015002005\u003c/span\u003e\u003cspan address=\"10.1177/0268580900015002005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSahebi, S., \u0026amp; Formosa, P. (2024). Artificial Intelligence (AI) and Global Justice. \u003cem\u003eMinds and Machines\u003c/em\u003e, \u003cem\u003e35\u003c/em\u003e(1), 4. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s11023-024-09708-7\u003c/span\u003e\u003cspan address=\"10.1007/s11023-024-09708-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShams, R. A., Zowghi, D., \u0026amp; Bano, M. (2023). AI and the quest for diversity and inclusion: a systematic literature review. \u003cem\u003eAI and Ethics\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s43681-023-00362-w\u003c/span\u003e\u003cspan address=\"10.1007/s43681-023-00362-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStern, C., Jordan, Z., \u0026amp; Mcarthur, A. (2014). Developing the review question and inclusion criteria. \u003cem\u003eAmerican Journal of Nursing\u003c/em\u003e, \u003cem\u003e114\u003c/em\u003e(4), 53\u0026ndash;56. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1097/01.NAJ.0000445689.67800.86\u003c/span\u003e\u003cspan address=\"10.1097/01.NAJ.0000445689.67800.86\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUNESCO. (2021). AI and education: guidance for policy-makers. In \u003cem\u003eAI and education: guidance for policy-makers\u003c/em\u003e. \u003cem\u003eUNESCO Publishing\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.54675/pcsp7350\u003c/span\u003e\u003cspan address=\"10.54675/pcsp7350\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUpadhyay, U., Gradisek, A., Iqbal, U., Dhar, E., Li, Y. C., \u0026amp; Syed-Abdul, S. (2023). Call for the responsible artificial intelligence in the healthcare. \u003cem\u003eBMJ Health \u0026amp; Care Informatics\u003c/em\u003e, \u003cem\u003e30\u003c/em\u003e(1), e100920. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/bmjhci-2023-100920\u003c/span\u003e\u003cspan address=\"10.1136/bmjhci-2023-100920\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVetter, M. A., Lucia, B., Jiang, J., \u0026amp; Othman, M. (2024). Towards a framework for local interrogation of AI ethics: A case study on text generators, academic integrity, and composing with ChatGPT. \u003cem\u003eComputers and Composition\u003c/em\u003e, \u003cem\u003e71\u003c/em\u003e, 102831. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.compcom.2024.102831\u003c/span\u003e\u003cspan address=\"10.1016/j.compcom.2024.102831\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWakunuma, K., \u0026amp; Eke, D. (2024). Africa, ChatGPT, and Generative AI Systems: Ethical Benefits, Concerns, and the Need for Governance. \u003cem\u003ePhilosophies\u003c/em\u003e, \u003cem\u003e9\u003c/em\u003e(3). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/philosophies9030080\u003c/span\u003e\u003cspan address=\"10.3390/philosophies9030080\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWiese (2025). \u003cem\u003eAI Ethics in Education Systematic Literature Review\u003c/em\u003e. Grail (Governance \u0026amp; Responsible AI LAB). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.grail-lab.org/publications/ai-ethics-in-education-systematic-literature-review?utm_source=chatgpt.com\u003c/span\u003e\u003cspan address=\"https://www.grail-lab.org/publications/ai-ethics-in-education-systematic-literature-review?utm_source=chatgpt.com\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYadav, R., Yadav, M., Huzooree, G., Najeeb, A., \u0026amp; Ranasinghe, A. (2025). Accessibility and inclusivity in AI-powered personalized learning. \u003cem\u003eTransformative AI Practices for Personalized Learning Strategies\u003c/em\u003e, 47\u0026ndash;67. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4018/979-8-3693-8744-3.ch003\u003c/span\u003e\u003cspan address=\"10.4018/979-8-3693-8744-3.ch003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYilma, K. (2025). Ethics of AI in Africa: Interrogating the role of Ubuntu and AI governance initiatives. \u003cem\u003eEthics and Information Technology\u003c/em\u003e, \u003cem\u003e27\u003c/em\u003e(2), 24. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10676-025-09834-5\u003c/span\u003e\u003cspan address=\"10.1007/s10676-025-09834-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZajko, M. (2021). Conservative AI and social inequality: conceptualizing alternatives to bias through social theory. \u003cem\u003eAI \u0026amp; SOCIETY\u003c/em\u003e, \u003cem\u003e36\u003c/em\u003e(3), 1047\u0026ndash;1056. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00146-021-01153-9\u003c/span\u003e\u003cspan address=\"10.1007/s00146-021-01153-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZembylas, M. (2023). A decolonial approach to AI in higher education teaching and learning: strategies for undoing the ethics of digital neocolonialism. \u003cem\u003eLearning Media and Technology\u003c/em\u003e, \u003cem\u003e48\u003c/em\u003e(1), 25\u0026ndash;37. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/17439884.2021.2010094\u003c/span\u003e\u003cspan address=\"10.1080/17439884.2021.2010094\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"AI ethics in education, Decolonial governance, Glocalized AI ethics, Epistemic justice, Participatory AI governance, Global South perspectives, Educational artificial intelligence, Systematic review","lastPublishedDoi":"10.21203/rs.3.rs-8417921/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8417921/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe rapid integration of artificial intelligence (AI) into educational systems has intensified global efforts to establish ethical guidelines governing its use. However, prevailing AI ethics frameworks in education remain predominantly shaped by Global North epistemologies, universalist moral assumptions, and centralized governance models. As a result, ethical principles frequently fail to translate into contextually legitimate or enforceable practices, particularly in Global South and postcolonial educational settings. This study addresses this gap by critically examining how AI ethics in education is produced, governed, and operationalized across diverse contexts.\u003c/p\u003e \u003cp\u003eUsing a PRISMA-guided systematic review of 84 peer-reviewed studies published between 2015 and 2025, this study employs an integrated \u003cb\u003ePICo\u0026ndash;Thematic synthesis\u003c/b\u003e to examine populations, interests, and contexts that are often marginalized in global AI ethics discourse. The analysis reveals three core findings: (1) AI ethics frameworks in education are heavily centralized in Global North institutions, with limited participatory governance and persistent algorithmic bias; (2) pluriversal and localized ethical practices\u0026mdash;grounded in Indigenous knowledge systems, communal ethics, and culturally embedded pedagogies\u0026mdash;have emerged as viable counter-models; and (3) ethical effectiveness is empirically associated with governance mechanisms that embed ethics into institutional participation, procurement, and accountability structures rather than voluntary principle adoption.\u003c/p\u003e \u003cp\u003eBuilding on these findings, the study advances the \u003cb\u003eFramework for Glocalized AI Ethics in Education (FGAIEE)\u003c/b\u003e as its central theoretical contribution. FGAIEE reconceptualizes AI ethics as a multi-scalar governance process that integrates epistemic pluriversality, participatory oversight, and structural enforceability. Rather than proposing another universal ethics model, the framework enables ethical coordination between global AI infrastructures and locally articulated educational values.\u003c/p\u003e \u003cp\u003eThis study contributes to scholarship on AI ethics, education, and decolonial governance by translating critical theory into an operational framework with global relevance. By repositioning ethics as an issue of epistemic justice and institutional design, FGAIEE offers policymakers, educators, and researchers a pathway to move AI ethics in education from symbolic compliance toward structural redress.\u003c/p\u003e","manuscriptTitle":"Decolonizing AI Ethics in Education: A Systematic Review and the Framework for Glocalized AI Ethics in Education (FGAIEE)","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-25 03:08:02","doi":"10.21203/rs.3.rs-8417921/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"0bc364f6-d933-410b-a049-38261ab2f95b","owner":[],"postedDate":"December 25th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-03-19T08:26:23+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-25 03:08:02","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8417921","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8417921","identity":"rs-8417921","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}