Inclusion and Equity in Artificial Intelligence (AI): Analysis of Educational Policies in Austria, Germany, and Switzerland

Inclusion and Equity in Artificial Intelligence (AI): Analysis of Educational Policies in Austria, Germany, and Switzerland | 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 Inclusion and Equity in Artificial Intelligence (AI): Analysis of Educational Policies in Austria, Germany, and Switzerland Edvina Bešić, Lea Schulz, Christa Schmid-Meier This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7232992/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 This paper examines how inclusion and educational equity are conceptualized and thematized in national AI and education policies in Germany, Austria, and Switzerland (the DACH region). Through a qualitative content analysis of 10 policy documents, this study explores how these policies articulate inclusion and equity, and whether they include concrete implementation plans. The analysis was conducted using MAXQDA, combining inductive codes related to inclusion and equity with a deductive framework. The policy analysis reveals a contrast in how the DACH region approaches inclusive and equitable AI education. Germany uses explicit equity metrics and funding, Austria relies on implicit STEM-focused inclusion, and Switzerland’s decentralized model yields fragmented, declarative measures, all three lacking enforceable commitments or measurable targets. This paper presents the first comparative analysis of how inclusion and equity are addressed in AI policy documents in the DACH region. It highlights promising examples of inclusive governance while also pinpointing structural barriers that may impede equitable access to AI-supported learning. Building on these insights, this paper develops targeted recommendations for future policy development, advocating not only for the integration of inclusion and equity as core principles in national AI policies but also for conceiving inclusion as a collaborative, cross-border effort that leverages shared good practices and joint oversight. Artificial Intelligence Policies Inclusive Education Educational Equity DACH region Introduction Artificial Intelligence (AI) is rapidly transforming key societal domains, including education. AI is commonly defined as a machine-based system that can generate outputs, such as predictions, content, or decisions, that influence physical or virtual environments (OECD, 2024). It is widely seen as holding great promise for transforming learning by enhancing students’ experiences both inside and outside the classroom, as well as affecting their physical, social-emotional, and intellectual development (Varsik & Vosberg, 2024; Salas-Pilco, 2020). AI-driven tools, such as intelligent tutoring systems, adaptive learning platforms, and chatbots, are praised for their potential to provide personalization and improve accessibility (Melo-López et al., 2025). At the same time, as AI technologies are increasingly embedded into educational systems, concerns about inclusion, equity, accessibility, and systemic bias are becoming more urgent (Shams, Zowghi & Bano, 2025). Inclusion needs to be understood as a process that responds to the diversity of learners by increasing participation and decreasing exclusion within and from education. It involves modifying content, approaches, and structures, based on the conviction that all children of suitable age should be educated within the regular system (UNESCO, 2005). Grounded in principles of equal opportunity, human rights, and social justice, inclusive education aims to ensure that all learners, regardless of ability, background, or identity, have access to meaningful and high-quality learning experiences (Ainscow, 2020). In this regard, the intersection of AI and inclusion offers both opportunities and challenges. As stated, AI holds significant potential to expand access to education and provide personalized support (Varsik & Vosberg, 2024). At the same time, if not intentionally designed and managed, it may reinforce or even worsen existing inequalities (ibid.). Without policy frameworks that actively promote inclusion and equity, AI risks maintaining normative learner profiles and marginalizing those who fall outside dominant assumptions (Shams, Zowghi, & Bano, 2025). Therefore, the absence of comprehensive ethical governance in education can lead to unintended consequences. Insufficient regulation may expose students to risks such as data misuse, algorithmic bias, and academic dishonesty. Educational institutions are thus at a critical crossroads, balancing the promise of innovation with its ethical and legal challenges (Ghimire & Edwards, 2024). Yet, despite increasing attention to AI in education over the past four decades (Dillenbourg, 2016; Kent & Du Boulay, 2022), the intersection with inclusion and equity as overarching principles remains underexplored (Varsik & Vosberg, 2024). Most studies focus on technological innovation or pedagogical effectiveness, while few critically examine how inclusion and equity are conceptualized and operationalized within these initiatives (ibid.). This gap becomes especially salient in light of international policy frameworks that explicitly call for inclusive digital transformation. Both the OECD and UNESCO emphasize the importance of using AI to foster inclusion and equity in education, while also warning of the risks posed by algorithmic bias and unequal access (OECD, 2024; Varsik & Vosberg, 2024). In the absence of clear equity measures, the benefits of AI may disproportionately accrue to privileged groups, further widening the educational divide (van Dijk, 2020). To date, over 50 governments worldwide have published national AI strategies that have set forth their aims and approaches to AI research, development, and deployment (Dua et al., 2025). These strategies often address multiple sectors, including education, and offer valuable insights into how governments perceive AI’s role in shaping future societies. However, as previous research has shown, few of these documents meaningfully engage with inclusion and equity-related dimensions (Schiff, 2022). In this context, analyzing how inclusion and equity are conceptualized and operationalized within national education and AI policies has become an urgent task. This article addresses that need by examining how inclusion and equity are addressed in national education and AI policies from Germany, Austria, and Switzerland (the DACH region). These three countries share linguistic, cultural, and political similarities but differ in the structure and ambition of their digital and educational strategic agendas. By comparing 10 official documents, this study aims to identify whether and how inclusion and equity-related principles are framed, which governance mechanisms are proposed, and what this implies for inclusive and equitable AI implementation in education. The analysis is grounded in an inclusion and equity-oriented perspective (UNESCO, 2017) and draws on qualitative content analysis to systematically code and compare the selected documents. In doing so, this paper contributes to a better understanding of the AI policy landscape in the DACH region. It provides differentiated recommendations for aligning AI development with the goal of inclusive and equitable education for all. Theoretical Framework Conceptual Clarifications Inclusion has become a key normative concept in international educational discourse rooted in the principles of human rights, social justice, and equity. UNESCO (2005, p. 13) defines inclusion as follows: Inclusion is seen as a process of addressing and responding to the diversity of needs of all learners through increasing participation in learning, cultures and communities, and reducing exclusion within and from education. It involves changes and modifications in content, approaches, structures and strategies, with a common vision which covers all children of the appropriate age range and a conviction that it is the responsibility of the regular system to educate all children. This definition also aligns with Ainscow’s (2020) understanding of inclusion, which we used in this paper. Namely, that inclusion is a process, “concerned with the identification and removal of barriers to the presence, participation, and achievement of all students” (p. 9). Moreover while considering all learners, a specific emphasis needs to be given to “those groups of learners who may be at risk of marginalization, exclusion or underachievement” (p. 9). Inclusion then challenges deficit-based approaches and calls for a systemic transformation in educational design and delivery. Zowghi and da Rimini (2023) emphasize the participatory aspect of Ainscow’s (2020) definition, arguing that in the AI context, inclusion entails active engagement and representation of diverse stakeholders. Achieving inclusive education, particularly in the context of utilizing AI in education, depends heavily on educational equity as a vital element. Equity focuses on ensuring fair access and participation by addressing systemic barriers and inequalities, thereby affirming that learners’ education is seen as equally important (UNESCO, 2017). In the realm of AI systems, equity refers to the fair distribution of benefits and the reduction of potential harm through targeted measures across the entire AI lifecycle. While inclusion in the AI context aims at equal participation of all users, equity requires proactive interventions and specific data strategies to ensure fair outcomes for all population groups, particularly historically marginalized communities (World Economic Forum, 2022). The connection between inclusive education and related concepts such as equity, participation, and access highlights its multidimensional nature (Kozleski, Artiles, & Waitoller, 2014). However, the absence of a shared conceptual understanding, combined with the heterogeneous nature of AI applications, hampers coherent discourse, research, and implementation. In the context of digital transformation, inclusive principles remain essential for critically assessing potentials and risks of AI in education. Artificial Intelligence (AI), as defined by the OECD, refers to a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. Different AI systems vary in their levels of autonomy and adaptiveness after deployment (OECD, 2024, p. 4). AI encompasses a broad set of technologies, including machine learning, natural language processing, and computer vision. In education, these technologies are increasingly used to support personalized learning, provide real-time feedback, and automate administrative processes (Shams, Zowghi, & Bano, 2025). The growing presence of AI in classrooms raises new demands on learners, not only in terms of usage but also in terms of understanding, assessing, and shaping these technologies. In this context, the OECD (2025, p. 4) describes AI literacy as the combination of “technical knowledge, durable skills, and future-ready attitudes” needed to succeed in an AI-influenced world. “It enables learners to engage, create with, manage, and design AI, while critically evaluating its benefits, risks, and ethical implications” (OECD, 2025, p. 4). This definition emphasizes critical, participatory, and ethical engagement, further highlighting the need for inclusive frameworks in AI education. Yet, the concept of AI literacy itself remains contested and definitions differ across countries and organizations. Intersections of AI and Inclusion The intersection of AI and inclusion opens new possibilities for designing learning environments that respond more effectively to diverse learners (Schmid-Meier & Schulz, 2025). Proponents have highlighted the potential of AI in enhancing accessibility. For instance, speech recognition, real-time translation, and personalized feedback systems (Shams, Zowghi & Bano, 2025; Varsik & Vosberg, 2024). AI-based tools can also support differentiation and individualization, enabling teachers and systems to adapt instruction based on learners’ prior knowledge, pace, language preferences, and support needs (Melo-López et al., 2025). This aligns with key principles of inclusive pedagogy. Inclusive pedagogy prioritizes flexibility, responsiveness, and a learner-centered design, which are well-supported by AI technologies (Ahmed et al., 2025; Jackson-Summers et al., 2024). Furthermore, the integration of AI into educational settings can facilitate collaborative learning and promote metacognitive skills, contributing to a constructivist and inclusive learning environment (Wahono, 2025). Moreover, AI has been used to develop assistive technologies that support students with disabilities, for example, by enabling alternative forms of communication, facilitating navigation, or compensating for sensory impairments (Arias-Flores et al., 2025). These developments have the potential to expand participation and enable more flexible forms of engagement, particularly for learners facing structural or physical barriers in conventional educational settings. However, these opportunities are associated with substantial risk. Without careful design, AI systems may inadvertently reproduce or even intensify existing educational inequalities (Varsik & Vosberg, 2024; van Dijk, 2020). A key issue is algorithmic discrimination, which can occur when AI systems generate outcomes that systematically disadvantage specific groups. This is often linked to biases in training data, where historical inequalities are encoded and perpetuated (Varsik & Vosberg, 2024). Additionally, many AI systems rely on culturally embedded assumptions about what constitutes a “normal” or “ideal” learner, assumptions that may exclude students whose identities, experiences, or learning styles do not align with dominant norms (Shams, Zowghi, & Bano, 2025). The lack of representativeness in datasets, limited transparency in algorithmic decision-making, and low accountability mechanisms all contribute to unequal outcomes. As Ghimire and Edwards (2024) warn, inadequate policy frameworks governing the ethical use of AI in education can expose students to risks such as data misuse, algorithmic bias, and academic dishonesty. Another challenge lies in the limited AI literacy among educators, school leaders, and policymakers. AI literacy involves not only the ability to use AI-powered tools but also to critically understand how they work, their limitations, and their broader implications. Without this understanding, there is a risk that AI tools will be adopted uncritically or misused, worsening educational disparities rather than reducing them (Mouta, Pinto-Llorente & Torrecilla-Sánchez, 2024). Therefore, the effective and equitable adoption of these tools requires continuous teacher training and awareness to address ethical and sociopolitical considerations effectively. Teachers should be able to approach these technologies with an understanding of their potential limitations and ethical challenges (Jackson-Summers et al., 2024). Taken together, these risks underscore the importance of integrating inclusion and equity principles into both the design and governance of AI technologies. If AI is to support, rather than undermine, inclusion and equity, its implementation must be accompanied by systemic reflection on who is included in designing them, how learners are represented, and what forms of participation are enabled or constrained by algorithmic systems. Methods The Research Context The development and implementation of AI policies in education are closely linked to national governance structures and political agenda. In the DACH region, AI policies are shaped by varying levels of decentralization and institutional responsibility. In Germany, digital education policy is shaped by federalism, where the Länder hold exclusive responsibility for education (GG Art. 30, 70). At the same time, the Standing Conference of the Ministers of Education and Cultural Affairs (KMK) harmonizes standards and frameworks. Complementary federal funding programs, formerly managed by the BMBF and by the BMFSFJ since 2025, foster innovation, digital infrastructure, media‑pedagogical concepts, and the development of digital competencies among teachers and learners. Governance is more centralized in Austria. The Austrian Government nationally coordinated different action plans for the country's digital transformation. The Federal Ministry of Education plays a key role in shaping education and digitalization policies. Switzerland differs significantly from Austria and Germany, as it operates a highly decentralized education system in which responsibilities are divided among the cantons and, to some extent, the municipalities. Unlike Austria and Germany, as EU member states, Switzerland currently lacks a dedicated AI legislation (Federal Office of Communications, 2025). Initiatives related to digital education have remained fragmented across cantonal jurisdictions. Objectives and Research Question This analysis examined how the nexus between AI, inclusion, and equity is conceptualized and thematized in national education and AI policies in the DACH region. Specifically, this review addresses the following questions: Q1. How are inclusion and equity addressed and conceptualized within education and AI policies? Q2. How do national education and AI policies in the DACH region align or differ in their treatment of the relationship between AI, inclusion, and equity? Document Selection Criteria To analyze the relationship between artificial intelligence (AI) and inclusive education and equity within the DACH region, clear inclusion and exclusion criteria were applied to guide the document selection process. The inclusion criteria were as follows: (i) documents had to explicitly address the integration of AI in educational settings, either partially or fully; (ii) only documents with credible authorship, issued or endorsed by national or regional ministries of education or official government agencies, were included; and (iii) documents had to be publicly available at the time of data collection, ensuring their relevance and timeliness. Documents were excluded if they lacked substantive references to AI, or were unrelated to education. The selection of countries and subnational entities within the DACH region was informed by both practical considerations and the aim of analyzing diverse yet comparable approaches to AI integration in education. Due to the overwhelming volume of German AI‑in‑education documents across 16 federal states compared to Switzerland and Austria, only two were selected as exemplars (Schleswig‑Holstein and Mecklenburg‑Vorpommern). Both published comprehensive, publicly accessible policy papers and shared similar legal frameworks, enabling a focused, in‑depth analysis within the project’s time and resource constraints. In Switzerland, policy documents from the Canton of Zurich were included. As the largest canton in Switzerland, Zurich was chosen because it has taken a leading role in AI-in-education policy development, providing accessible guidance with both legal and pedagogical significance. Additionally, a document from the Digitalization Working Group of the Conference of Cantonal Education Departments for Compulsory Schooling in the German-speaking Cantons of Switzerland (DVK) was selected because it serves as an inter-cantonal document, especially in the absence of a comprehensive national framework. In the Austrian context, selection was straightforward because the federal government has an official national AI policy. Additionally, the Federal Ministry of Education published a handout and launched an initiative (KI-Initiative) specifically focused on AI in education. These were included in this analysis as they represent the Austrian government's official efforts toward AI in education, providing primary data on national goals, policies, and implementation. The 10 documents selected are listed in Table 1 below. In this paper, we refer to each document using numbers listed in this table. Table 1 Overview of Selected National AI and Education Policy Documents No. Country Title Year Publisher Purpose 1 AT Federal Government Strategy for Artificial Intelligence. Artificial Intelligence Mission Austria 2030 (AIM AT 2030) 2021 Federal Ministry for Climate Action, Environment, Energy, Mobility, Innovation and Technology (BMK); Federal Ministry for Digital and Economic Affairs (BMDW) Outlines Austria’s national vision, strategic goals, and policy measures for the development, deployment, and governance of artificial intelligence across sectors, including education 2 AT Annex to AIM AT 2030 2021 BMK; BMDW Serves as a concrete action plan that defines initial, sector-specific applications of artificial intelligence (e.g., in education, mobility, energy, agriculture) and outlines practical implementation measures and pilot projects 3 AT Engagement with Artificial Intelligence in the Education System 2023 Federal Ministry of Education, Science and Research (BMBWF) Offers guidance, ethical considerations, and practical tools for educators and schools 4 DE Recommendations for Education Administration on Handling Artificial Intelligence in Educational Processes 2024 Secretariat of the Standing Conference of Ministers of Education and Cultural Affairs Guides state education authorities in the selection, deployment, and quality assurance of AI applications in teaching and administration 5 DE Large Language Models and Their Potential in the Education System. Position Paper by the SWK 2024 Standing Scientific Commission of the Standing Conference (SWK) Assesses pedagogical opportunities and risks of LLMs, formulates criteria for use in teaching, and provides best-practice scenarios and evidence-based recommendations 6 DE Guideline – Exploring the World of Generative AI Systems Together 2023 Ministry of Education and Childcare of Mecklenburg-Western Pomerania Introduces schools to generative AI systems via practical examples, templates, and ethical guidance 7 DE AI@School – Tips for Initial Guidance for Schools 2023 Ministry of General and Vocational Education, Science, Research and Culture (MBWFK), Schleswig-Holstein Provides schools with guidance on piloting AI tools, infrastructure, training, and didactic use 8 CH Artificial Intelligence in Compulsory Schooling 2025 Department of Education, Canton of Zurich Equips schools in Zurich with orientation for responsible, equitable, and legally compliant AI use in education 9 CH Artificial Intelligence in Education – Legal Best Practices 2025 Department of Education, Canton of Zurich; Innovation Zurich; Zurich Metropolitan Conference Summarizes key federal and cantonal regulations for AI and provides compliance guidance for schools and solution providers 10 CH Questions on the Use of AI in Compulsory Schooling 2025 Digitalization Working Group of the Conference of Cantonal Education Departments for Compulsory Schooling in the German-Speaking Cantons of Switzerland (DVK) Compiles questions on AI in education, clarifies responsibilities and lays the foundat Analytical Method The documents were analyzed in their original language, which is German, and direct quotations in this paper are translations provided by the authors. The analysis process involved an initial screening of documents, followed by a content analysis using a structured, multi-stage approach to qualitative content analysis, as developed by Kuckartz and Rädiker (2024). The three authors independently coded all documents using a predefined coding schema based on the theoretical framework discussed in this paper. We created a codebook (see Table 2 and Table 3) containing code definitions, examples, and subcodes to ensure consistency in coding and thematic interpretation. Text data were divided into discrete units, such as sentences or paragraphs, with each unit independently assigned to one or more relevant codes. Discrepancies were reviewed and resolved through a consensus discussion. After an initial discussion, the coding scheme was refined slightly, with clarifications of code boundaries and the merging of overlapping categories, while remaining true to the original framework. All coding was conducted using MAXQDA, which recorded coder decisions, facilitating the integration of frequency counts and cross-case comparisons. Table 2 Code Book: Theoretical Constructs of Inclusion and Equity for AI Code Definition Coding Example Subcodes 1.1 Concepts Use of terms such as inclusion and related concepts explicitly in the context of AI and schooling. AI can help to foster inclusion within schools. • Inclusion • Participation • Participatory input • Diversity • Heterogeneity • Educational justice • Equity of opportunity 1.2 Target Groups References to specific groups intended to benefit from inclusion efforts, particularly vulnerable or marginalized students (e.g., students with disabilities, refugee or migrant students, students with learning difficulties, or students from non-dominant linguistic/cultural backgrounds). Students with special educational needs can be individually supported through adaptive AI tools. 1.3 Potentials Statements describing the opportunities AI offers to support inclusion in education. Focus on educational benefits and inclusive design. AI enables accessibility of learning materials. • Differentiation & Individualization (e.g., subtitles, screen readers) • Collaborative work & learning • Assessment (feedback, diagnostics, learning progress) • Gamification • Social presence (e.g., AI bots or robots) • Shared learning content • Addressing heterogeneity 1.4 Exclusion Risks Statements about barriers or disadvantages related to AI, especially those affecting vulnerable groups. Considers both direct and indirect impacts. Not all students have access to AI-based learning systems. AI systems must also be available in plain language. • Technological barriers (access, digital divide) • Linguistic and cultural barriers • Bias and distortions (e.g., stereotypes, gender bias) • Cognitive barriers (text complexity, overload) • Socioeconomic factors (equipment, home support) • Discrimination (implicit or explicit) 1.5 Systemic Conditions for Inclusive Use of AI in Schools Describes framework conditions that enable or hinder inclusive AI application in schools. Includes infrastructure, training, access, and competence. Teachers need training in the use of AI-supported tools for students with visual impairments. Not all schools have the necessary digital infrastructure. • Infrastructure and Equipment (devices, tools, assistive technology, internet access) • Teacher Training (focus on inclusion; pre-service and in-service AI training) • Economic / Access Factors (inequality in access to devices and tools; availability of AI tools at home) • AI Literacy Among Students (ability to use and critically assess AI) To answer Q3, we included the coding of more overarching categories related to educational policy and normative frameworks relevant to inclusive education, as shown in Table 3 below. Table 3 Code‑Book: Strategies and normative frameworks for Inclusion & Equity Results The following chapter offers a structured analysis of national AI policy papers from the DACH region, focusing on how inclusion and equity principles are conceptualized and operationalized within them. Terminology The analysis of terminology related to the “inclusive context” – meaning the frequency of selected key terms within the respective national policies (see Table 4) – showed that concepts such as equity, educational and participatory justice, or diversity were only marginally represented in the analyzed documents. Germany had the highest count, with a total of 12 mentions, followed by Switzerland (n = 6), and Austria (n = 4). However, these frequencies primarily serve to illustrate how individual policy documents engage with inclusion-related terms, rather than directly comparing the three countries. Differences in the number and length of documents across the countries limit their comparability and should be considered when interpreting these findings. The most frequently used term was equity (Chancengerechtigkeit), particularly in German (n = 4) and Swiss (n = 3) documents. Other terms, such as inclusion (Inklusion), participation (Teilhabe), and educational justice (Bildungsgerechtigkeit), appeared in German documents, and to a lesser extent, in those from Austria and Switzerland. In contrast, terms such as participatory involvement (Partizipation), diversity (Diversität), and heterogeneity (Heterogenität) were mentioned only sporadically or not at all. Table 4 Frequency of Terms Related to the “Inclusive Context” in the AI Policies of the DACH Countries Codes Germany Austria Switzerland Equity of opportunity 4 1 3 Educational justice 2 0 2 Inclusion 2 1 0 Participation 2 1 1 Participatory input 1 0 0 Heterogeneity 1 0 0 Diversity 0 1 0 Target Groups Across the analyzed documents, references to target groups varied between the three countries in terms of specificity and inclusivity. German documents featured the most differentiated terminology, explicitly addressing a range of specific student groups. These included terms such as “all learners” (mentioned four times), “weaker learners,” “high-performing or low-performing students,” “heterogeneous student body,” “persons with disabilities,” and “disadvantaged” or “marginalized groups.” The language ranged from general references to education-wide inclusion to more narrowly defined groups with particular needs. In contrast, Austrian documents relied more heavily on broad, generic designations. While groups such as “disadvantaged groups,” “learners across the entire educational pathway,” and “women and girls” were mentioned, the framing tended to align with overarching societal goals, such as promoting education and participation. There was little differentiation between specific learner needs, and the descriptions remained general, focusing on the population as a whole rather than on distinct subgroups. Swiss documents have also adopted a predominantly universalist approach. Most references to learners were broad, such as “at all school levels” and “regardless of origin, socioeconomic status, or physical and cognitive abilities.” Only a few more specific mentions appeared, such as references to “dyslexia and dyscalculia” and “particularly sensitive personal data” including “health data.” Similar to the Austrian documents, the Swiss documents did not provide detailed differentiation of individual groups, instead emphasizing inclusivity through general, all-encompassing language. Systemic Conditions When analyzing the systemic condition, the German documents once again led the comparison to Austria and Switzerland, especially in terms of code frequency and their interpretation. The main focus across all countries in this theme was AI literacy. The importance of AI literacy was frequently emphasized in German documents (n = 37), which also referred to economic and access factors, stating it as a prerequisite for reducing the digital divide. One representative statement for this is: “A basic understanding of how the technology functions, as well as the ability to recognize the opportunities, limitations, and risks of new digital possibilities, is essential for a responsible and reflective use, as well as for promoting learning in the classroom and at home” (see document number 4, p. 6). Additionally, in several passages, concerns were expressed about equitable economic access: “The first-level divide described above must be addressed through appropriate measures. The federal states therefore aim to provide a joint interface solution that enables data-protection-compliant and cost-free access to LLMs in the school sector in the near future” (4, p. 14). However, the initial and ongoing training of teachers was discussed only once in the context of inclusion and equity. In Austria, AI literacy was mentioned less often (n = 5). Statements such as: “The goal is the responsible and pedagogically/didactically meaningful use of AI-based tools as support for teachers and learners, always with critical reflection on the framework conditions and possible consequences, such as data protection (personal data)” (1, p. 22) suggest a stronger emphasis on learning with AI rather than learning about AI. Neither access to devices nor the professional teacher development were explicitly addressed as a systemic condition. This suggests that the structural framework requirements remain largely underrepresented in Austrian documents. In the selected Swiss documents, AI literacy was also addressed less often than Germany (n = 6). One example stated: “Students must learn to engage competently and responsibly with this reality and its consequences (AI literacy). Children and adolescents should be supported in primary school on their way to a mature and informed use of generative ML functions” (10, p. 10). Additionally, there was one mention each of technical infrastructure and teacher training, noting: “Teacher educators at universities of teacher education continue their professional development and can demonstrate in both pre-service and in-service training how generative ML functions can be used for diagnostics and support planning” (10, p. 8). AI’s Exclusion Risks When discussing the potential risks that AI may pose in education, the German AI policy papers identified numerous exclusion risks, and Germany was the only one among the three countries that addressed all subcategories of this theme, showing a broader awareness of exclusion issues. For example, one passage stated: “It is problematic that LLMs are not equally accessible to all learners. Barriers to access may arise from limited language proficiency, dyslexia, or visual impairments. Learners from low socio-economic backgrounds may face financial obstacles due to private providers and lack of support from parents” (5, p. 17). Bias and distortion (n = 8) were among the most frequently addressed issues, with specific focus on topics such as: “A phenomenon closely related to hallucinations in the functioning of large language models (LLMs) is bias—systematic distortions of information (Navigli et al., 2023), which arise from the selection of training data (Gombert et al., 2023). These biases could influence a wide range of dimensions, including moral orientation (Schramowski et al., 2022), religion, gender, ethnicity, profession (Nadeem et al., 2021), or political alignment (Feng et al., 2023; Rozado, 2023). Different LLMs may, for instance, adopt different political stances (Feng et al., 2023) or reproduce prejudices (e.g., Bergener et al., 2023)“ (6, p. 9). Technological barriers were also discussed (n = 6), identifying the digital divide as a major challenge for educational equity. Cognitive barriers, including linguistic and cultural challenges, were mentioned in three segments, along with socio-economic factors. Discrimination and linguistic-cultural barriers were addressed only once. Comparably to Germany, the Austrian documents frequently addressed bias and distortions (n = 6) as well as discrimination (n = 5). An exemplary segment highlighting discrimination related to marginalized groups reads as follows: “At the same time, it is important to prevent discrimination, support disadvantaged groups, and ensure diversity” (1, p. 31). However, these perspectives often remain normative and less concrete regarding their implementation. Other barriers, such as technological (n = 3) and socioeconomic aspects (n = 1), were only marginally represented. Cognitive and linguistic-cultural barriers have not been addressed. This suggests that structural and infrastructural conditions for access received comparatively limited attention in the Austrian documents. In contrast to Germany and Austria, Swiss documents displayed a somewhat different emphasis regarding exclusion risks. Technological barriers (n = 6) were most frequently addressed, highlighting functional and infrastructural aspects. Bias and distortions (n = 4) were also prominently mentioned, for instance: “The risks of AI must always be considered (distortions, sources, etc.)” (8, p. 6). Socioeconomic aspects were moderately represented (n = 3), while discrimination, cognitive barriers, and linguistic-cultural barriers appeared only once. However, it is important to note that these latter three dimensions were often mentioned together in a single coded segment, potentially distorting individual frequency counts: “The schools guarantee that all children have access to the new opportunities with AI. Regardless of origin, socioeconomic status, and physical or cognitive abilities, all children should be able to participate equally” (8, p. 2). AI’s Inclusive Potentials When discussing the potential risks of AI in education, it is essential to equally consider its potential benefits. The analysis of German documents revealed a clear emphasis on AI’s capability for differentiation, individualization, and adaptivity (n = 51), highlighting its use primarily as a tool to tailor learning experiences to individual learner needs, including technical accessibility aspects such as subtitles. For example, the Secretariat of the Standing Conference of Ministers of Education and Cultural Affairs explicitly states: “They also assume that adaptive and AI-supported learning materials can have a positive effect on the acquisition of basic competencies” (4, pp. 6–7). Additionally, assessment-related potentials such as feedback provision, diagnostics, and learning analytics were prominently acknowledged (n = 20). Cooperative learning (n = 8) and managing heterogeneity (n = 3) received some attention, whereas gamification was only marginally referenced (n = 2). Notably, the social presence of AI bots or robots was entirely absent (n = 0), suggesting a predominantly functional-technical rather than a socially interactive framing of AI’s educational potential. In the Austrian documents, the primary focus is on differentiation, individualization, and adaptivity (n = 19), emphasizing AI's potential for tailoring learning experiences, including considerations of technical accessibility such as subtitles. Additionally, assessment-related potentials, such as feedback provision, diagnostics, and learning analytics, were mentioned (n = 6). Addressing heterogeneity within classrooms received some attention (n = 4), reflecting an awareness of educational diversity and associated challenges. In contrast, the social presence of AI bots or robots was mentioned only marginally (n = 1). Cooperative learning and gamification potentials were entirely absent. Swiss documents showed limited, yet focused, engagement with AI’s educational potentials. Differentiation, individualization, and adaptivity were mentioned three times (n = 3), indicating an understanding of AI as a means for personalized learning. assessment-related aspects such as feedback and diagnostics were also addressed (n = 2), alongside references to managing heterogeneity (n = 2), suggesting some recognition of diverse learner needs. However, no studies have been found regarding cooperative learning, gamification, or the social presence of AI bots or robots. This absence reinforces the impression that Swiss documents primarily frame AI as a technical and individualized support tool, with little emphasis on its interactive or socially embedded dimensions. Normative and Participatory Frameworks While analyzing the policy documents, we also aimed to determine whether underlying inclusive frameworks were present. The analysis revealed that the AI policy documents from the DACH region neither explicitly address inclusion and equity nor reference international frameworks such as the UN Convention on the Rights of Persons with Disabilities (UN CRPD) or the Sustainable Development Goals (SDGs) related to inclusive education. Furthermore, only four passages were found that addressed participation and co-determination, understood as the involvement of students, persons with disabilities, or other affected groups in decision-making or design processes. Three of these references appear in the Secretariat of the Standing Conference of the Ministers of Education and Cultural Affairs policies in Germany (4). Among other aspects, they highlighted that informational self-determination is a fundamental right of children and adolescents that must be respected when using AI-based systems. One passage addressed the involvement of teachers and students in selecting AI applications. Another passage emphasized participation as an important educational goal. The fourth reference, appearing in the Austrian AIM document (1), stats that AI development should be human-centered and that participation should be a core design principle. Although the publisher, the Federal Ministry of Innovation, Mobility, and Infrastructure Republic Austria, states that over 160 experts from different fields (i.e. technology, economics, natural sciences, law, social sciences, and education) were involved in writing the paper, it’s unclear who participated or if they have expertise in inclusive education. However, none of the documents provided a detailed explanation or operationalization of the concept of participation. Discussion The comparative analysis of the selected documents reveals a consistent gap between rhetorical commitments and policies for inclusion and equity in the AI field across the DACH countries. While Germany shows broader engagement, particularly in terms of diverse target group mentions, system-level conditions, and risk awareness, its policy documents still fall short of embedding inclusion as a binding governance principle. Austria and Switzerland demonstrate even less thematic depth, often relying on declarative, generalized statements without concrete mechanisms. These patterns reflect what Varsik and Vosberg ( 2024 ) describe as the “superficial engagement” with inclusive digital transformation, highlighting the disconnect between AI's transformative potential and the ongoing marginalization of inclusion and equity as policy priorities. Although Germany uses terms related to inclusion and equity more frequently than Austria and Switzerland do, their overall occurrence remains low. This suggests that inclusion and equity are only marginally addressed in the AI policy documents of all three DACH countries. Although Germany nominally leads with 12 mentions of inclusion and equity-related key terms, this picture becomes more nuanced when considering the significantly greater length and scope of its policy documents. This highlights a structural pattern of marginal visibility of inclusion and equity across the DACH region. It is important to note that these findings are based solely on frequency analysis, without assessing qualitative differences in thematic depth. For instance, the policy paper from the Standing Conference of the Ministers of Education and Cultural Affairs (KMK) in Germany dedicated an entire chapter to equity, illustrating a structural focus on the issue. Such a detailed treatment is absent in Austrian or Swiss documents. These findings should therefore be viewed as initial indicators of the theme’s visibility. The cross-country comparison reveals noticeable differences in how target groups are addressed in the AI policy papers of Germany, Austria, and Switzerland. Germany demonstrates the widest range of formulations, suggesting a more nuanced understanding of educational inequalities and a more targeted approach to vulnerable groups, although the language used can sometimes be vague or lack systematic operationalization. In contrast, the Austrian policy papers emphasize broader target groups. While there are references to “disadvantaged groups” and gender-specific support goals (“women and girls”), the overall language remains mainly declarative. The mention of “learners across the entire educational pathway” indicates a systemic understanding of education but lacks detailed reference to varying support needs. While all three DACH countries express general commitments to inclusion in their AI-in-education policies, only Germany shows tentative efforts toward a differentiated approach to target groups. Austria and especially Switzerland rely predominantly on universal formulations, often referring to the entire student population without specifying particular needs. This raises questions about the depth and effectiveness of their inclusion claims. Drawing on the work of Ainscow ( 2020 ), inclusive education is understood not merely as a universal provision for all students but as a dynamic process that also requires a targeted emphasis on those groups at risk of marginalization, exclusion, or underachievement. Since inclusive education is founded on the belief that education is a fundamental human right and a cornerstone for a more just society, it is crucial to explicitly name and address the needs of vulnerable groups. Without such specificity, policies may overlook the unique challenges these groups face, thereby diluting the equity and fairness that inclusive policies aim to achieve. Following Ainscow’s ( 2020 ) framework, inclusive education entails not only the physical presence of all learners but also their active participation and achievement. Most AI policies, however, primarily focus on presence, with little attention to systemic participation or tailored support mechanisms. This limited conceptualization also neglects the form of ‘representational participation’ emphasized by Zowghi and da Rimini ( 2023 ), where inclusion requires the active involvement and visibility of diverse learners in both the development and deployment of AI systems. From this perspective, the lack of specific references to vulnerable groups in Austrian and Swiss documents may reflect a gap between stated inclusive ideals and the operationalization of those ideals in policy. Naming and addressing specific groups, such as learners with disabilities, socioeconomically disadvantaged students, or those with learning differences, is essential for ensuring that equity is not merely aspirational but actionable within AI-driven educational transformations, as stated by Shams, Zowghi, and Bano ( 2025 ). Germany addresses systemic conditions most comprehensively, especially AI literacy. While Switzerland and Austria also touch on this topic, their coverage is less extensive. Notably, Germany recognizes the importance of infrastructure and teacher training as relevant components for effective AI integration into education, whereas these aspects are nearly absent in Austrian documents. The near-complete absence of explicit teacher training programs in the Austrian and Swiss documents reveals structural capacity gaps in transferring AI literacy into pedagogical practice. The comparison reveals that systemic requirements are prioritized differently across the region, with Germany taking a more proactive stance. These disparities matter, as teacher training is widely regarded as essential for equitable adaptation of AI in education. Without an understanding of AI, there is a risk of uncritical use that may worsen education inequities (Mouta, Pinto-Liorente, & Torrecilla-Sánchez, 2024). Continued training that addresses ethical and sociopolitical issues is crucial, enabling teachers to engage with AI in a critical and responsible manner (Jackson-Summers et al., 2024 ). Germany also demonstrated the most comprehensive engagement with exclusion risks. All seven subcodes were coded at least once, indicating a nuanced awareness of the issue. Austria takes a more selective approach, mainly focusing on discrimination and stereotypical bias, while neglecting technical access issues. While Austria focuses heavily on the debate around algorithmic bias, systemic dimensions, such as infrastructural inequalities, remain largely peripheral. In contrast, Switzerland concentrates primarily on technical barriers, with social and intersectional perspectives receiving little attention. Compared to Germany, the thematic range in Switzerland appears to be more limited, but it is clearly focused on learning-oriented adaptation. As Shams, Zowghi, and Bano ( 2025 ) emphasize, AI systems that are developed without critical reflection on their data foundations and design logic risk not only replicating but reinforcing existing inequalities. In the analyzed policies, however, such structural forms of bias are rarely addressed explicitly or accompanied by concrete countermeasures. Switzerland finds a middle ground: various barriers are acknowledged, but with limited depth. Overall, the codings suggest that Swiss documents frame exclusion risks predominantly through a technological lens, with comparatively less focus on social or intersectional perspectives. As Shams, Zowghi, and Bano ( 2025 ) argue, unexamined training data and implicit design assumptions can cause the systemic reproduction of educational inequalities. Representative justice, ensuring that diverse learners are visible and considered in AI development, is not meaningfully addressed in any of the reviewed documents. This absence reinforces the impression that Swiss documents primarily frame AI as a technical and individualized support tool, with little emphasis on its interactive or socially embedded dimensions. The complete omission of socially interactive AI scenarios, such as robotics applications designed to foster social-emotional competencies, highlights a narrow, instrumental-functional AI discourse that underestimates relational learning processes. This suggests significant divergence in national perspectives on inclusion-related challenges in AI: Germany adopts a more systemic approach, Austria emphasizes social justice in a narrower sense, and Switzerland mainly concentrates on linguistic and technological aspects. The analysis shows that Germany most frequently identified potential uses of AI, especially in support areas such as “differentiation, individualization/adaptivity,” which were coded 51 times. Austria explores similar potentials but less frequently. Notably, neither Switzerland nor Austria mentions cooperative learning, although it appears multiple times in German documents. Gamification remains a marginal topic across all three countries. Overall, these findings suggest that AI’s potential is mainly understood in functional terms, however, only Germany reflects a broader range of didactic opportunities particularly focusing on GenAI. Possibilities such as diagnostics with AI or augmented reality were not mentioned. Only Germany briefly mentioned participatory principles, including references to informational self-determination and stakeholder involvement. Austria and Switzerland remain entirely declarative or silent on this matter. Across all three countries, references to international frameworks such as the UN CRPD or SDGs are notably absent. The disregard for international reference frameworks undermines the stated ambition of aligning national AI policies with global goals for inclusion. Overall, participation and normative frameworks are weakly developed and mostly symbolic, rather than actionable. This aligns with Armstrong, Armstrong, and Spandagou’s ( 2011 ) critique that inclusion often remains symbolic rather than structural, as reflected in the analyzed documents, which exhibit a rhetorical rather than operational commitment to participation and co-determination. From an ‘equity-by-design’ perspective (WEF, 2022), effective inclusion demands not only generalized commitments but the deliberate implementation of design principles that reflect the needs and identities of historically marginalized groups. The absence of representational justice in these documents, manifested in the lack of explicit group references and inclusive governance mechanisms, limits the transformative potential of AI as outlined in UNESCO’s and WEF’s global frameworks (UNESCO, 2017 ; WEF, 2022). Given this context, the findings emphasize the need to integrate inclusion and participation into AI policy explicitly and structurally. As outlined in the theoretical framework, inclusion involves more than access; it requires representational justice, differentiated support, and participatory governance (UNESCO, 2005 ; Armstrong et al., 2011 ). The near absence of references to international equity frameworks, such as the UN CRPD or SDGs, presents a missed opportunity to align national policies with global commitments to educational justice. Transitioning from symbolic gestures to binding actions is crucial to ensure that AI development genuinely promotes inclusive and equitable learning environments. Conclusion and Implications for Policy, Education, Research In light of the preceding analysis, the following recommendations should be viewed not as supplementary measures but as an essential framework for fostering inclusive digital policy. The analysis of AI policy documents across the DACH countries reveals significant gaps in the integration of inclusion and equity principles into national AI strategic papers. Although Germany addresses AI literacy more extensively than Austria and Switzerland, this focus often remains detached from concrete, inclusive measures or differentiated implementation for marginalized groups. This lack of specificity poses challenges, as genuinely inclusive education requires explicitly targeting vulnerable populations and actively dismantling structural barriers (Ainscow, 2020 ). Germany`s policies identify a broader range of exclusion risks and potentials of AI, including individualized and cooperative learning methods. In contrast, Austrian and Swiss documents predominantly frame AI’s potential through personalization and assessment, with limited acknowledgment of associated risks such as technological and socioeconomic barriers. The overarching emphasis on AI as a solution for individualized learning, particularly in German and Austrian documents, risks reinforcing a techno-solutionist narrative (Mochizuky et al., 2025) that overlooks structural inequalities and the need for accessible, participatory digital learning environments (Autenrieth et al., 2025). To avoid the “techno-ableism” trap, where AI tools are assumed to be neutral or universally effective, future policies or recommendations should be based on intersectional perspectives and co-designed with inclusion experts (Shew, 2020 ). As emphasized by Autenrieth et al. (2025), transformative AI education requires more than just technological infrastructure; it needs ethical reflection, normative considerations, and participatory governance to promote educational justice. Before stating the recommendations of this paper, it should be noted that all three countries notably omit international equity frameworks, such as the UN CRPD and the SDGs, thereby missing an opportunity to align with global commitments to educational justice (UNESCO, 2017 ). For future policy and research, several recommendations have emerged from the results. First, inclusion must be established as a binding principle within national AI policies guided explicitly by international frameworks (UN-CRPD, SDGs). Second, systematic identification and targeted support for vulnerable learner groups should become standard practice. Third, inclusive AI competencies should be embedded comprehensively in teacher education to foster critical, equitable AI use (Autenrieth et al., 2025). Fourth, while the German documents provide isolated examples of good practices, they lack systematic evaluation. A research agenda focusing on the evaluation of these within real school settings should be supported by national policy funding. Finally, participatory governance mechanisms involving different stakeholders, such as educators, inclusion experts, and affected communities should be strengthened (Shams, Zowghi, & Bano, 2025 ). This approach demands a shift from techno-centric to democratic and inclusive AI governance models, which is crucial for realizing the transformative potential of AI in education. In line with UNESCO’s ( 2005 ) call for a systemic transformation of educational systems and Autenrieth et al.’s (2025) vision of a ‘democratic logic’ in AI education, future policy frameworks must be grounded in inclusive governance, equity-by-design, and representational justice. To conclude, the document analysis revealed that inclusion and equity principles are not yet firmly embedded in AI educational policy papers. To move beyond symbolic rhetoric, it is essential to openly address participation, non-discrimination, and diversity, integrating these principles into both educational design and regulatory frameworks (UNICEF, 2017). This aligns with Autenrieth et al.`s (2025) call for a shift from an “innovation logic” to a “democratic logic” in AI governance in education. Achieving truly inclusive AI education requires more than rhetorical alignment; it demands a comprehensive shift from innovation-centric models to participatory, justice-based policies and strategies. Limitations of the Study This study has several limitations. First, it focuses only on official policies, excluding non-official documents such as reports from NGOs and educational institutions. This means that it does not capture how the policies are implemented in practice or their impact on educational equity and inclusion. Second, the analysis includes documents published between 2021 and May 2025, which may miss recent updates or drafts, especially those related to developments around the EU AI Act. Third, the framework used for inclusion and equity reflects dominant Western perspectives, and the qualitative coding process involves some subjectivity, despite efforts to mitigate this through independent coding and consensus discussions. Future research could improve this by triangulating the findings with interviews or focus groups with policymakers or other stakeholders to validate document interpretations. Fourth, this study’s geographical and cultural context is also a limitation. While the DACH region shares some cultural and political similarities, there are important differences that could affect how AI and inclusion are implemented and understood in education. Future research might include examining sub-national differences within the DACH region. Declarations Funding The authors declare that no funds, grants, or other support were received during the preparation of this manuscript. Competing Interests The authors have no relevant financial or non-financial interests to disclose. Author Contributions All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by Edvina Bešić, Christa Schmid-Meier, and Lea Schulz. The first draft of the manuscript's introduction was written by Christ Schmid-Meier, the Methods section by Edvina Bešić, and the analysis by Lea Schulz and Edvina Bešić. The other parts of the paper were written in collaboration by the three authors, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. References Ahmed, S., Rahman, M. S., Kaiser, M. S., & Hosen, A. S. M. S. (2025). Advancing Personalized and Inclusive Education for Students with Disability Through Artificial Intelligence: Perspectives, Challenges, and Opportunities. Digital , 5 (2), 11. https://doi.org/10.3390/digital5020011 Ainscow, M. (2020). Promoting inclusion and equity in education: lessons from international experiences. Nordic Journal of Studies in Educational Policy , 6 (1), 7–16. https://doi.org/10.1080/20020317.2020.1729587 Armstrong, D., Armstrong, A. C., & Spandagou, I. (2011). Inclusion: By choice or by chance? International Journal of Inclusive Education , 15 (1), 29–39. https://doi.org/10.1080/13603116.2010.496192 Autenrieth, D., Schluchter, J. R., & Schulz, L. (in press). AI is all you need? Zeitschrift für Inklusion. Arias-Flores, H., Valencia-Aragón, K., Calle-Jimenez, T., & Sanchez-Gordon, S. (2025). Artificial Intelligence and Assistive Technologies: A Systematic Review of Educational Applications for Disabilities. In M. Antona & C. Stephanidis (Eds.), Universal Access in Human-Computer Interaction (Vol. 15780, pp. 283–292). Springer Nature Switzerland. https://doi.org/10.1007/978-3-031-93848-1_19 Dillenbourg, P. (2016). The Evolution of Research on Digital Education. International Journal of Artificial Intelligence in Education , 26 (2), 544–560. https://doi.org/10.1007/s40593-016-0106-z Dong, M., Conway, J. R., Bonnefon, J. F., Shariff, A., & Rahwan, I. (2024). Fears about artificial intelligence across 20 countries and six domains of application. American Psychologist . https://doi.org/10.1037/amp0001454 Dua, M., Singh, J. P., & Shehu, A. (2025). The ethics of national artificial intelligence plans: an empirical lens. AI and Ethics , 5 (4). https://doi.org/10.1007/s43681-025-00663-2 Federal Office of Communications (2025). Artificial Intelligence: Overview and Switzerland’s regulatory approach. Artificial Intelligence . https://www.bakom.admin.ch/bakom/en/homepage/digital-switzerland-and-internet/strategie-digitale-schweiz/ai.html Ghimire, A., & Edwards, J. (2024). From Guidelines to Governance: A Study of AI Policies in Education. In A. M. Olney, I.-A. Chounta, Z. Liu, O. C. Santos, & I. I. Bittencourt (Eds.), Artificial Intelligence in Education. Posters and Late Breaking Results, Workshops and Tutorials, Industry and Innovation Tracks, Practitioners, Doctoral Consortium and Blue Sky (Vol. 2151, pp. 299–307). Springer Nature Switzerland. https://doi.org/10.1007/978-3-031-64312-5_36 Jackson-Summers, A. G., Mrakovcich, K. L., Gray, J. P., Fleischmann, C. M., Emami, T., & Page, E. J. (2024). A systematic review of inclusive pedagogical research using the CIRTL inclusive pedagogy framework: multi-disciplinary and STEM perspectives, current trends and a research agenda. Discover Education , 3 (1), 30. https://doi.org/10.1007/s44217-024-00093-y Kent, C., & Boulay, B. D. (2022). AI for Learning (1st ed.). CRC. https://doi.org/10.1201/9781003194545 Kozleski, E., Artiles, A., & Waitoller, F. (2014). Equity in inclusive education: A cultural historical comparative perspective. In The SAGE handbook of special education (Vol. 2, pp. 231–245). SAGE. https://doi.org/10.4135/9781446282236.n16 Kuckartz, U., & Rädiker, S. (2024). Qualitative Inhaltsanalyse. Methoden, Praxis, Umsetzung mit Software und künstlicher Intelligenz (6. Auflage) [Qualitative Content Analysis: Methods, Practice, Implementation with Software and Artificial Intelligence]. Juventa Verlag. Melo-López, V. A., Basantes-Andrade, A., Gudiño-Mejía, C. B., & Hernández-Martínez, E. (2025). The Impact of Artificial Intelligence on Inclusive Education: A Systematic Review. Education Sciences , 15 (5), 539. https://doi.org/10.3390/educsci15050539 Mochizuki, Y., Bruillard, E., & Bryan, A. (2025). The ethics of AI or techno-solutionism? UNESCO’s policy guidance on AI in education. British Journal of Sociology of Education , 1–22. https://doi.org/10.1080/01425692.2025.2502808 Monett, D., & Paquet, G. (2025). Against the Commodification of Education — if harms then not AI. Journal of Open Distance and Digital Education , 2 (1). https://doi.org/10.25619/WAZGW457 Mouta, A., Pinto-Llorente, A. M., & Torrecilla-Sánchez, E. M. (2024). Uncovering Blind Spots in Education Ethics: Insights from a Systematic Literature Review on Artificial Intelligence in Education. International Journal of Artificial Intelligence in Education , 34 (3), 1166–1205. https://doi.org/10.1007/s40593-023-00384-9 OECD (2024). Explanatory memorandum on the updated OECD definition of an AI system (OECD Artificial Intelligence Papers No. 8). https://doi.org/10.1787/623da898-en OECD (2025). Empowering learners for the age of AI: An AI literacy framework for primary and secondary education (Review draft) . https://ailiteracyframework.org/?utm_source=substack&utm_medium=email Salas-Pilco, S. Z. (2020). The impact of AI and robotics on physical, social‐emotional and intellectual learning outcomes: An integrated analytical framework. British Journal of Educational Technology , 51 (5), 1808–1825. https://doi.org/10.1111/bjet.12984 Schiff, D. (2022). Education for AI, not AI for Education: The Role of Education and Ethics in National AI Policy Strategies. International Journal of Artificial Intelligence in Education , 32 (3), 527–563. https://doi.org/10.1007/s40593-021-00270-2 Schmid-Meier, C., & Schulz, L. (2025). Exploring the Inclusive Potential of AI Chatbots: Engagement and Barriers in Foreign Language Training. In M. Antona & C. Stephanidis (Eds.), Universal Access in Human-Computer Interaction (Vol. 15780, pp. 377–395). Springer Nature Switzerland. https://doi.org/10.1007/978-3-031-93848-1_25 Shams, R. A., Zowghi, D., & Bano, M. (2025). AI and the quest for diversity and inclusion: a systematic literature review. AI and Ethics , 5 (1), 411–438. https://doi.org/10.1007/s43681-023-00362-w Shew, A. Ableism, Technoableism, and, & Future, A. I. (2020). IEEE Technology and Society Magazine , 39 (1), 40–85. https://doi.org/10.1109/mts.2020.2967492 UNESCO (2005). Guidelines for Inclusion: Ensuring Access to Education for All. UNESCO. UNESCO. (2017). A Guide for ensuring inclusion and equity in education . UNESCO. van Dijk, J. (2020). The Digital Divide . Polity. Varsik, S., & Vosberg, L. (2024). The potential impact of Artificial Intelligence on equity and inclusion in education (OECD Artificial Intelligence Papers No. 23). https://doi.org/10.1787/15df715b-en Wahono, Z. (2025). Integrating AI into Constructivist Pedagogy: Strategies and Outcomes. Advances in Nonlinear Variational Inequalities , 28 (5s), 227–231. https://doi.org/10.52783/anvi.v28.3789 Zowghi, D., & da Rimini, F. (2023). Diversity and Inclusion in Artificial Intelligence. arXiv preprint arXiv:2305.12728, 2023. World Economic Forum [WEF] (2022). A Blueprint for Equity and Inclusion in Artificial Intelligence . (n.d.). Retrieved July 19, 2025, from https://www3.weforum.org/docs/WEF_A_Blueprint_for_Equity_and_Inclusion_in_Artificial_Intelligence_2022.pdf Additional Declarations No competing interests reported. Supplementary Files AppendixA.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-7232992","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":492582565,"identity":"f53de803-6f43-499a-8662-27f4e958ec9c","order_by":0,"name":"Edvina Bešić","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABCElEQVRIie3QsUoDMRzH8X8InEtK1hwt+ArpKLTcq1wodLJFKDjdcOUgXYJd7zH6Bkb+cF1E1xtcDsGpw0FBuijGQ8QiEbo55DsELtyHHwQgFPqXpWC7E0gOLQADegoh5U/C/iDfJz36y0f42eUQrzKYc47L/Vg/DeSWWjhkcJ54SGx2EssKFnGpiv5MvzCJUUpMBUPjIbKepsgiUJuaaDrTyOKCSejlQHwk6cg7qNvHO00vOsJb8pZD4l0RE4s97VZAaUoc4e4RqFtRPiLunx25EQtRqyI2D58kkjioxMRYz4utVLFnr6M5X2+b9nCNScSxaXbZaLzKPTNfY8ef9tdNKBQKhU7qA5bEVYcWVae+AAAAAElFTkSuQmCC","orcid":"","institution":"University of Vienna","correspondingAuthor":true,"prefix":"","firstName":"Edvina","middleName":"","lastName":"Bešić","suffix":""},{"id":492582566,"identity":"bf496ce2-c760-44e8-9535-794ac819a3f8","order_by":1,"name":"Lea Schulz","email":"","orcid":"","institution":"University of Flensburg","correspondingAuthor":false,"prefix":"","firstName":"Lea","middleName":"","lastName":"Schulz","suffix":""},{"id":492582567,"identity":"11742455-51e3-48f0-9799-02ccef27628d","order_by":2,"name":"Christa Schmid-Meier","email":"","orcid":"","institution":"University of Teacher Education in Special Needs","correspondingAuthor":false,"prefix":"","firstName":"Christa","middleName":"","lastName":"Schmid-Meier","suffix":""}],"badges":[],"createdAt":"2025-07-28 10:53:14","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7232992/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7232992/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":93622795,"identity":"d4513a7f-4c14-40b9-9ac1-8ef00159aa35","added_by":"auto","created_at":"2025-10-15 18:17:12","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":83585,"visible":true,"origin":"","legend":"","description":"","filename":"ArticleforReviewDACHAIInclusionEquity.docx","url":"https://assets-eu.researchsquare.com/files/rs-7232992/v1/bd22fb2d3b7e460aaad95135.docx"},{"id":93622291,"identity":"1cddd245-3cb0-41b8-a22a-735219d6a73a","added_by":"auto","created_at":"2025-10-15 18:09:12","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5453,"visible":true,"origin":"","legend":"","description":"","filename":"92c34177898a471bb677a4e097c89881.json","url":"https://assets-eu.researchsquare.com/files/rs-7232992/v1/0f56ddc83ace3f27aa89c823.json"},{"id":93622293,"identity":"dc2d9f03-1c6e-4d5f-8534-5cda6f7bc4fc","added_by":"auto","created_at":"2025-10-15 18:09:12","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":130293,"visible":true,"origin":"","legend":"","description":"","filename":"92c34177898a471bb677a4e097c898811enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7232992/v1/3be8091e26da865daa164a04.xml"},{"id":93622294,"identity":"10cc0828-9d81-4719-85b5-7324c030f94a","added_by":"auto","created_at":"2025-10-15 18:09:12","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":128943,"visible":true,"origin":"","legend":"","description":"","filename":"92c34177898a471bb677a4e097c898811structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7232992/v1/1cf2c47fbd21dac3e7047488.xml"},{"id":93622796,"identity":"67d99da7-fce0-4c6b-9f56-12622c0b3833","added_by":"auto","created_at":"2025-10-15 18:17:12","extension":"html","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":136376,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7232992/v1/5bf4f43776ea24fd6e8bc99d.html"},{"id":104959853,"identity":"807d678d-eb3a-4d81-b7ab-c1b4c3c20764","added_by":"auto","created_at":"2026-03-19 08:43:52","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":711473,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7232992/v1/04080216-fe99-4295-81d1-b58903e09d18.pdf"},{"id":93622290,"identity":"fd4714d7-bcf3-47b4-bfb2-32b150f0d281","added_by":"auto","created_at":"2025-10-15 18:09:12","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":16253,"visible":true,"origin":"","legend":"","description":"","filename":"AppendixA.docx","url":"https://assets-eu.researchsquare.com/files/rs-7232992/v1/47fd0cb0750a09ba25e32c14.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Inclusion and Equity in Artificial Intelligence (AI): Analysis of Educational Policies in Austria, Germany, and Switzerland","fulltext":[{"header":"Introduction","content":"\u003cp\u003eArtificial Intelligence (AI) is rapidly transforming key societal domains, including education. AI is commonly defined as a machine-based system that can generate outputs, such as predictions, content, or decisions, that influence physical or virtual environments (OECD, 2024). It is widely seen as holding great promise for transforming learning by enhancing students\u0026rsquo; experiences both inside and outside the classroom, as well as affecting their physical, social-emotional, and intellectual development (Varsik \u0026amp; Vosberg, 2024; Salas-Pilco, 2020). AI-driven tools, such as intelligent tutoring systems, adaptive learning platforms, and chatbots, are praised for their potential to provide personalization and improve accessibility (Melo-L\u0026oacute;pez et al., 2025). At the same time, as AI technologies are increasingly embedded into educational systems, concerns about inclusion, equity, accessibility, and systemic bias are becoming more urgent (Shams, Zowghi \u0026amp; Bano, 2025).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eInclusion needs to be understood as a process that responds to the diversity of learners by increasing participation and decreasing exclusion within and from education. It involves modifying content, approaches, and structures, based on the conviction that all children of suitable age should be educated within the regular system (UNESCO, 2005). Grounded in principles of equal opportunity, human rights, and social justice, inclusive education aims to ensure that all learners, regardless of ability, background, or identity, have access to meaningful and high-quality learning experiences\u0026nbsp;(Ainscow, 2020).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn this regard, the intersection of AI and inclusion offers both opportunities and challenges. As stated, AI holds significant potential to expand access to education and provide personalized support (Varsik \u0026amp; Vosberg, 2024). At the same time, if not intentionally designed and managed, it may reinforce or even worsen existing inequalities (ibid.). Without policy frameworks that actively promote inclusion and equity, AI risks maintaining normative learner profiles and marginalizing those who fall outside dominant assumptions (Shams, Zowghi, \u0026amp; Bano, 2025). Therefore, the absence of comprehensive ethical governance in education can lead to unintended consequences. Insufficient regulation may expose students to risks such as data misuse, algorithmic bias, and academic dishonesty. Educational institutions are thus at a critical crossroads, balancing the promise of innovation with its ethical and legal challenges (Ghimire \u0026amp; Edwards, 2024).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eYet, despite increasing attention to AI in education over the past four decades (Dillenbourg, 2016; Kent \u0026amp; Du Boulay, 2022), the intersection with inclusion and equity as overarching principles remains underexplored (Varsik \u0026amp; Vosberg, 2024). Most studies focus on technological innovation or pedagogical effectiveness, while few critically examine how inclusion and equity are conceptualized and operationalized within these initiatives (ibid.). This gap becomes especially salient in light of international policy frameworks that explicitly call for inclusive digital transformation. Both the OECD and UNESCO emphasize the importance of using AI to foster inclusion and equity in education, while also warning of the risks posed by algorithmic bias and unequal access (OECD, 2024; Varsik \u0026amp; Vosberg, 2024). In the absence of clear equity measures, the benefits of AI may disproportionately accrue to privileged groups, further widening the educational divide (van Dijk, 2020).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo date, over 50 governments worldwide have published national AI strategies that have set forth their aims and approaches to AI research, development, and deployment (Dua et al., 2025). These strategies often address multiple sectors, including education, and offer valuable insights into how governments perceive AI\u0026rsquo;s role in shaping future societies. However, as previous research has shown, few of these documents meaningfully engage with inclusion and equity-related dimensions (Schiff, 2022). In this context, analyzing how inclusion and equity are conceptualized and operationalized within national education and AI policies has become an urgent task.\u003c/p\u003e\n\u003cp\u003eThis article addresses that need by examining how inclusion and equity are addressed in national education and AI policies from Germany, Austria, and Switzerland (the DACH region). These three countries share linguistic, cultural, and political similarities but differ in the structure and ambition of their digital and educational strategic\u0026nbsp;agendas. By comparing 10 official documents, this study aims to identify whether and how inclusion and equity-related principles are framed, which governance mechanisms are proposed, and what this implies for inclusive and equitable AI implementation in education.\u003c/p\u003e\n\u003cp\u003eThe analysis is grounded in an inclusion and equity-oriented perspective (UNESCO, 2017) and draws on qualitative content analysis to systematically code and compare the selected documents. In doing so, this paper contributes to a better understanding of the AI policy landscape in the DACH region. It provides differentiated recommendations for aligning AI development with the goal of inclusive and equitable education for all.\u003c/p\u003e\n\u003ch2\u003eTheoretical Framework\u0026nbsp;\u003c/h2\u003e\n\u003ch2\u003eConceptual Clarifications\u003c/h2\u003e\n\u003cp\u003eInclusion has become a key normative concept in international educational discourse rooted in the principles of human rights, social justice, and equity.\u0026nbsp;UNESCO (2005, p. 13) defines inclusion as follows:\u003c/p\u003e\n\u003cp\u003eInclusion is seen as a process of addressing and responding to the diversity of needs of all learners through increasing participation in learning, cultures and communities, and reducing exclusion within and from education. It involves changes and modifications in content, approaches, structures and strategies, with a common vision which covers all children of the appropriate age range and a conviction that it is the responsibility of the regular system to educate all children.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThis definition also aligns with Ainscow\u0026rsquo;s (2020) understanding of inclusion, which we used in this paper. Namely, that inclusion is a process, \u0026ldquo;concerned with the identification and removal of barriers to the presence, participation, and achievement of all students\u0026rdquo; (p. 9). Moreover while considering all learners, a specific emphasis needs to be given to \u0026ldquo;those groups of learners who may be at risk of marginalization, exclusion or underachievement\u0026rdquo; (p. 9). Inclusion then challenges deficit-based approaches and calls for a systemic transformation in educational design and delivery. Zowghi\u0026nbsp;and\u0026nbsp;da\u0026nbsp;Rimini (2023) emphasize the participatory aspect of Ainscow\u0026rsquo;s (2020) definition, arguing that in the AI context, inclusion entails\u0026nbsp;active engagement and representation of diverse stakeholders.\u0026nbsp;Achieving inclusive education, particularly in the context of utilizing AI in education, depends heavily on educational equity as a vital element. Equity focuses on ensuring fair access and participation by addressing systemic barriers and inequalities, thereby affirming that learners\u0026rsquo; education is seen as equally important (UNESCO, 2017). In the realm of AI systems, equity refers to the fair distribution of benefits and the reduction of potential harm through targeted measures across the entire AI lifecycle. While inclusion in the AI context aims at equal participation of all users, equity requires proactive interventions and specific data strategies to ensure fair outcomes for all population groups, particularly historically marginalized communities (World Economic Forum, 2022).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe connection between inclusive education and related concepts such as equity, participation, and access highlights its multidimensional nature (Kozleski, Artiles, \u0026amp; Waitoller, 2014). However, the absence of a shared conceptual understanding, combined with the heterogeneous nature of AI applications, hampers coherent discourse, research, and implementation. In the context of digital transformation, inclusive principles remain essential for critically assessing potentials and risks of AI in education. Artificial Intelligence (AI), as defined by the OECD, refers to\u0026nbsp;\u003c/p\u003e\n\u003cp\u003ea machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. Different AI systems vary in their levels of autonomy and adaptiveness after deployment (OECD, 2024, p. 4).\u003c/p\u003e\n\u003cp\u003eAI encompasses a broad set of technologies, including machine learning, natural language processing, and computer vision. In education, these technologies are increasingly used to support personalized learning, provide real-time feedback, and automate administrative processes (Shams, Zowghi, \u0026amp; Bano, 2025).\u003c/p\u003e\n\u003cp\u003eThe growing presence of AI in classrooms raises new demands on learners, not only in terms of usage but also in terms of understanding, assessing, and shaping these technologies. In this context, the OECD (2025, p. 4) describes AI literacy as the combination of \u0026ldquo;technical knowledge, durable skills, and future-ready attitudes\u0026rdquo; needed to succeed in an AI-influenced world. \u0026ldquo;It enables learners to engage, create with, manage, and design AI, while critically evaluating its benefits, risks, and ethical implications\u0026rdquo; (OECD, 2025, p. 4). This definition emphasizes critical, participatory, and ethical engagement, further highlighting the need for inclusive frameworks in AI education. Yet, the concept of AI literacy itself remains contested and definitions differ across countries and organizations.\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eIntersections of AI and Inclusion\u003c/h2\u003e\n\u003cp\u003eThe intersection of AI and inclusion opens new possibilities for designing learning environments that respond more effectively to diverse learners (Schmid-Meier \u0026amp; Schulz, 2025). Proponents have highlighted the potential of AI in enhancing accessibility. For instance, speech recognition, real-time translation, and personalized feedback systems (Shams, Zowghi \u0026amp; Bano, 2025; Varsik \u0026amp; Vosberg, 2024). AI-based tools can also support differentiation and individualization, enabling teachers and systems to adapt instruction based on learners\u0026rsquo; prior knowledge, pace, language preferences, and support needs\u0026nbsp;(Melo-L\u0026oacute;pez et al., 2025). This aligns with key principles of inclusive pedagogy. Inclusive pedagogy prioritizes flexibility, responsiveness, and a learner-centered design, which are well-supported by AI technologies (Ahmed et al., 2025; Jackson-Summers et al., 2024). Furthermore, the integration of AI into educational settings can facilitate collaborative learning and promote metacognitive skills, contributing to a constructivist and inclusive learning environment (Wahono, 2025). Moreover, AI has been used to develop assistive technologies that support students with disabilities, for example, by enabling alternative forms of communication, facilitating navigation, or compensating for sensory impairments (Arias-Flores et al., 2025). These developments have the potential to expand participation and enable more flexible forms of engagement, particularly for learners facing structural or physical barriers in conventional educational settings.\u003c/p\u003e\n\u003cp\u003eHowever, these opportunities are associated with substantial risk. Without careful design, AI systems may inadvertently reproduce or even intensify existing educational inequalities (Varsik \u0026amp; Vosberg, 2024; van Dijk, 2020). A key issue is algorithmic discrimination, which can occur when AI systems generate outcomes that systematically disadvantage specific groups. This is often linked to biases in training data, where historical inequalities are encoded and perpetuated (Varsik \u0026amp; Vosberg, 2024). Additionally, many AI systems rely on culturally embedded assumptions about what constitutes a \u0026ldquo;normal\u0026rdquo; or \u0026ldquo;ideal\u0026rdquo; learner, assumptions that may exclude students whose identities, experiences, or learning styles do not align with dominant norms (Shams, Zowghi, \u0026amp; Bano, 2025). The lack of representativeness in datasets, limited transparency in algorithmic decision-making, and low accountability mechanisms all contribute to unequal outcomes. As Ghimire and Edwards (2024) warn, inadequate policy frameworks governing the ethical use of AI in education can expose students to risks such as data misuse, algorithmic bias, and academic dishonesty. Another challenge lies in the limited AI literacy among educators, school leaders, and policymakers. AI literacy involves not only the ability to use AI-powered tools but also to critically understand how they work, their limitations, and their broader implications. Without this understanding, there is a risk that AI tools will be adopted uncritically or misused, worsening educational disparities rather than reducing them (Mouta, Pinto-Llorente \u0026amp; Torrecilla-S\u0026aacute;nchez, 2024). Therefore, the effective and equitable adoption of these tools requires continuous teacher training and awareness to address ethical and sociopolitical considerations effectively. Teachers should be able to approach these technologies with an understanding of their potential limitations and ethical challenges (Jackson-Summers et al., 2024). Taken together, these risks underscore the importance of integrating inclusion and equity principles into both the design and governance of AI technologies. If AI is to support, rather than undermine, inclusion and equity, its implementation must be accompanied by systemic reflection on who is included in designing them, how learners are represented, and what forms of participation are enabled or constrained by algorithmic systems.\u003c/p\u003e"},{"header":"Methods","content":"\u003ch2\u003eThe Research Context\u0026nbsp;\u003c/h2\u003e\n\u003cp\u003eThe development and implementation of AI policies in education are closely linked to national governance structures and political agenda. In the DACH region, AI policies are shaped by varying levels of decentralization and institutional responsibility. In Germany, digital education policy is shaped by federalism, where the L\u0026auml;nder hold exclusive responsibility for education (GG Art. 30, 70). At the same time, the Standing Conference of the Ministers of Education and Cultural Affairs (KMK) harmonizes standards and frameworks. Complementary federal funding programs, formerly managed by the BMBF and by the BMFSFJ since 2025, foster innovation, digital infrastructure, media‑pedagogical concepts, and the development of digital competencies among teachers and learners. Governance is more centralized in Austria. The Austrian Government nationally coordinated different action plans for the country\u0026apos;s digital transformation. The Federal Ministry of Education plays a key role in shaping education and digitalization policies. Switzerland differs significantly from Austria and Germany, as it operates a highly decentralized education system in which responsibilities are divided among the cantons and, to some extent, the municipalities. Unlike Austria and Germany, as EU member states, Switzerland currently lacks a dedicated AI legislation (Federal Office of Communications, 2025). Initiatives related\u0026nbsp;to digital education have remained fragmented across cantonal jurisdictions.\u003c/p\u003e\n\u003ch2\u003eObjectives and Research Question\u003c/h2\u003e\n\u003cp\u003eThis analysis examined how the nexus between AI, inclusion, and equity is conceptualized and thematized in national education and AI policies in the DACH region. Specifically, this review addresses the following questions:\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003eQ1. How are inclusion and equity addressed and conceptualized within education and AI policies?\u003c/li\u003e\n \u003cli\u003eQ2. How do national education and AI policies in the DACH region align or differ in their treatment of the relationship between AI, inclusion, and equity?\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2\u003eDocument Selection Criteria\u003c/h2\u003e\n\u003cp\u003eTo analyze the relationship between artificial intelligence (AI) and inclusive education and equity within the DACH region, clear inclusion and exclusion criteria were applied to guide the document selection process. The inclusion criteria were as follows: (i) documents had to explicitly address the integration of AI in educational settings, either partially or fully; (ii) only documents with credible authorship, issued or endorsed by national or regional ministries of education or official government agencies, were included; and (iii) documents had to be publicly available at the time of data collection, ensuring their relevance and timeliness. Documents were excluded if they lacked substantive references to AI, or were unrelated to education.\u003c/p\u003e\n\u003cp\u003eThe selection of countries and subnational entities within the DACH region was informed by both practical considerations and the aim of analyzing diverse yet comparable approaches to AI integration in education. Due to the overwhelming volume of German AI‑in‑education documents across 16 federal states compared to Switzerland and Austria, only two were selected as exemplars (Schleswig‑Holstein and Mecklenburg‑Vorpommern). Both published comprehensive, publicly accessible policy papers and shared similar legal frameworks, enabling a focused, in‑depth analysis within the project\u0026rsquo;s time and resource constraints. In Switzerland, policy documents from the Canton of Zurich were included. As the largest canton in Switzerland, Zurich was chosen because it has taken a leading role in AI-in-education policy development, providing accessible guidance with both legal and pedagogical significance. Additionally, a document from the Digitalization Working Group of the Conference of Cantonal Education Departments for Compulsory Schooling in the German-speaking Cantons of Switzerland (DVK) was selected because it serves as an inter-cantonal document, especially in the absence of a comprehensive national framework. In the Austrian context, selection was straightforward because the federal government has an official national AI policy. Additionally, the Federal Ministry of Education published a handout and launched an initiative (KI-Initiative) specifically focused on AI in education. These were included in this analysis as they represent the Austrian government\u0026apos;s official efforts toward AI in education, providing primary data on national goals, policies, and implementation. The 10 documents selected are listed in Table 1 below. In this paper, we refer to each document using numbers listed in this table.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 1 Overview of Selected National AI and Education Policy Documents\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"3\" cellpadding=\"0\"\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eNo.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCountry\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eTitle\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eYear\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePublisher\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePurpose\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFederal Government Strategy for Artificial Intelligence. Artificial Intelligence Mission Austria 2030 (AIM AT 2030)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2021\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFederal Ministry for Climate Action, Environment, Energy, Mobility, Innovation and Technology (BMK); Federal Ministry for Digital and Economic Affairs (BMDW)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eOutlines Austria\u0026rsquo;s national vision, strategic goals, and policy measures for the development, deployment, and governance of artificial intelligence across sectors, including education\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAnnex to AIM AT 2030\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2021\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eBMK; BMDW\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eServes as a concrete action plan that defines initial, sector-specific applications of artificial intelligence (e.g., in education, mobility, energy, agriculture) and outlines practical implementation measures and pilot projects\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eEngagement with Artificial Intelligence in the Education System\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFederal Ministry of Education, Science and Research (BMBWF)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eOffers guidance, ethical considerations, and practical tools for educators and schools\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eRecommendations for Education Administration on Handling Artificial Intelligence in Educational Processes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2024\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eSecretariat of the Standing Conference of Ministers of Education and Cultural Affairs\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGuides state education authorities in the selection, deployment, and quality assurance of AI applications in teaching and administration\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eLarge Language Models and Their Potential in the Education System. Position Paper by the SWK\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2024\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eStanding Scientific Commission of the Standing Conference (SWK)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAssesses pedagogical opportunities and risks of LLMs, formulates criteria for use in teaching, and provides best-practice scenarios and evidence-based recommendations\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGuideline \u0026ndash; Exploring the World of Generative AI Systems Together\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eMinistry of Education and Childcare of Mecklenburg-Western Pomerania\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eIntroduces schools to generative AI systems via practical examples, templates, and ethical guidance\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAI@School \u0026ndash; Tips for Initial Guidance for Schools\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eMinistry of General and Vocational Education, Science, Research and Culture (MBWFK), Schleswig-Holstein\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eProvides schools with guidance on piloting AI tools, infrastructure, training, and didactic use\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eArtificial Intelligence in Compulsory Schooling\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2025\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDepartment of Education, Canton of Zurich\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eEquips schools in Zurich with orientation for responsible, equitable, and legally compliant AI use in education\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eArtificial Intelligence in Education \u0026ndash; Legal Best Practices\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2025\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDepartment of Education, Canton of Zurich; Innovation Zurich; Zurich Metropolitan Conference\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eSummarizes key federal and cantonal regulations for AI and provides compliance guidance for schools and solution providers\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eQuestions on the Use of AI in Compulsory Schooling\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2025\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDigitalization Working Group of the Conference of Cantonal Education Departments for Compulsory Schooling in the German-Speaking Cantons of Switzerland (DVK)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCompiles questions on AI in education, clarifies responsibilities and lays the foundat\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003ch2\u003eAnalytical Method\u003c/h2\u003e\n\u003cp\u003eThe documents were analyzed in their original language, which is German, and direct quotations in this paper are translations provided by the authors. The analysis process involved an initial screening of documents, followed by a content analysis using a structured, multi-stage approach to qualitative content analysis, as developed by Kuckartz and R\u0026auml;diker (2024). The three authors independently coded all documents using a predefined coding schema based on the theoretical framework discussed in this paper. We created a codebook (see Table 2 and Table 3) containing code definitions, examples, and subcodes to ensure consistency in coding and thematic interpretation. Text data were divided into discrete units, such as sentences or paragraphs, with each unit independently assigned to one or more relevant codes. Discrepancies were reviewed and resolved through a consensus discussion. After an initial discussion, the coding scheme was refined slightly, with clarifications of code boundaries and the merging of overlapping categories, while remaining true to the original framework. All coding was conducted using MAXQDA, which recorded coder decisions, facilitating the integration of frequency counts and cross-case comparisons.\u003c/p\u003e\n\u003cp\u003eTable 2 Code Book: Theoretical Constructs of Inclusion and Equity for AI\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"3\" cellpadding=\"0\"\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eCode\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDefinition\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCoding Example\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eSubcodes\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e1.1 Concepts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eUse of terms such as inclusion and related concepts explicitly in the context of AI and schooling.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAI can help to foster inclusion within schools.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026bull; Inclusion\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Participation\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Participatory input\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Diversity\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Heterogeneity\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Educational justice\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Equity of opportunity\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e1.2 Target Groups\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eReferences to specific groups intended to benefit from inclusion efforts, particularly vulnerable or marginalized students (e.g., students with disabilities, refugee or migrant students, students with learning difficulties, or students from non-dominant linguistic/cultural backgrounds).\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eStudents with special educational needs can be individually supported through\u0026nbsp;\u003c/p\u003e\n \u003cp\u003eadaptive AI tools.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e1.3 Potentials\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eStatements describing the opportunities AI offers to support inclusion in education. Focus on educational benefits and inclusive design.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAI enables accessibility of learning materials.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026bull; Differentiation \u0026amp; Individualization (e.g., subtitles, screen readers)\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Collaborative work \u0026amp; learning\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Assessment (feedback, diagnostics, learning progress)\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Gamification\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Social presence (e.g., AI bots or robots)\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Shared learning content\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Addressing heterogeneity\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e1.4 Exclusion Risks\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eStatements about barriers or disadvantages related to AI, especially those affecting vulnerable groups. Considers both direct and indirect impacts.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNot all students have access to AI-based learning systems.\u0026nbsp;\u003cbr\u003e\u0026nbsp;AI systems must also be available in plain language.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026bull; Technological barriers (access, digital divide)\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Linguistic and cultural barriers\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Bias and distortions (e.g., stereotypes, gender bias)\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Cognitive barriers (text complexity, overload)\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Socioeconomic factors (equipment, home support)\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Discrimination (implicit or explicit)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e1.5 Systemic Conditions for Inclusive Use of AI in Schools\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDescribes framework conditions that enable or hinder inclusive AI application in schools. Includes infrastructure, training, access, and competence.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eTeachers need training in the use of AI-supported tools for students with visual impairments.\u0026nbsp;\u003cbr\u003e\u0026nbsp;Not all schools have the necessary digital infrastructure.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026bull; Infrastructure and Equipment (devices, tools, assistive technology, internet access)\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Teacher Training (focus on inclusion; pre-service and in-service AI training)\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; Economic / Access Factors (inequality in access to devices and tools; availability of AI tools at home)\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u0026bull; AI Literacy Among Students (ability to use and critically assess AI)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo answer Q3, we included the coding of more overarching categories related to educational policy and normative frameworks relevant to inclusive education, as shown in Table 3 below.\u003c/p\u003e\n\u003cp\u003eTable 3 Code‑Book: Strategies and normative frameworks for Inclusion \u0026amp; Equity\u003c/p\u003e\n\u003cp\u003e\u003cimg src=\"https://myfiles.space/user_files/127393_c7e80a1c9bb65875/127393_custom_files/img1760432869.png\" style=\"width: 687px;\"\u003e\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eThe following chapter offers a structured analysis of national AI policy papers from the DACH region, focusing on how inclusion and equity\u0026nbsp;principles are conceptualized and operationalized within them.\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eTerminology\u003c/h2\u003e\n\u003cp\u003eThe analysis of terminology related to the \u0026ldquo;inclusive context\u0026rdquo; \u0026ndash; meaning the frequency of selected key terms within the respective national policies\u0026nbsp;(see Table 4) \u0026ndash; showed that concepts such as equity, educational and participatory justice, or diversity were only marginally represented in the analyzed documents. Germany had the highest count, with a total of 12 mentions, followed by Switzerland (n = 6), and Austria (n = 4). However, these frequencies primarily serve to illustrate how individual policy documents engage with inclusion-related terms, rather than directly comparing the three countries. Differences in the number and length of documents across the countries limit their comparability and should be considered when interpreting these findings. The most frequently used term was equity (Chancengerechtigkeit), particularly in German (n = 4) and Swiss (n = 3) documents. Other terms, such as inclusion (Inklusion), participation (Teilhabe), and educational justice (Bildungsgerechtigkeit), appeared in German documents, and to a lesser extent, in those from Austria and Switzerland. In contrast, terms such as participatory involvement (Partizipation), diversity (Diversit\u0026auml;t), and heterogeneity (Heterogenit\u0026auml;t) were mentioned only sporadically or not at all.\u003c/p\u003e\n\u003cp\u003eTable 4 Frequency of Terms Related to the \u0026ldquo;Inclusive Context\u0026rdquo; in the AI Policies of the DACH Countries\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"3\" cellpadding=\"0\"\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 205px;\"\u003e\n \u003cp\u003eCodes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003eGermany\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003eAustria\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003eSwitzerland\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 205px;\"\u003e\n \u003cp\u003eEquity of opportunity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 205px;\"\u003e\n \u003cp\u003eEducational justice\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 205px;\"\u003e\n \u003cp\u003eInclusion\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 205px;\"\u003e\n \u003cp\u003eParticipation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 205px;\"\u003e\n \u003cp\u003eParticipatory input\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 205px;\"\u003e\n \u003cp\u003eHeterogeneity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 205px;\"\u003e\n \u003cp\u003eDiversity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003ch2\u003eTarget Groups\u003c/h2\u003e\n\u003cp\u003eAcross the analyzed documents, references to target groups varied between the three countries in terms of specificity and inclusivity. German documents featured the most differentiated terminology, explicitly addressing a range of specific student groups. These included terms such as \u0026ldquo;all learners\u0026rdquo; (mentioned four times), \u0026ldquo;weaker learners,\u0026rdquo; \u0026ldquo;high-performing or low-performing students,\u0026rdquo; \u0026ldquo;heterogeneous student body,\u0026rdquo; \u0026ldquo;persons with disabilities,\u0026rdquo; and \u0026ldquo;disadvantaged\u0026rdquo; or \u0026ldquo;marginalized groups.\u0026rdquo; The language ranged from general references to education-wide inclusion to more narrowly defined groups with particular needs.\u003c/p\u003e\n\u003cp\u003eIn contrast, Austrian documents relied more heavily on broad, generic designations. While groups such as \u0026ldquo;disadvantaged groups,\u0026rdquo; \u0026ldquo;learners across the entire educational pathway,\u0026rdquo; and \u0026ldquo;women and girls\u0026rdquo; were mentioned, the framing tended to align with overarching societal goals, such as promoting education and participation. There was little differentiation between specific learner needs, and the descriptions remained general, focusing on the population as a whole rather than on distinct subgroups.\u003c/p\u003e\n\u003cp\u003eSwiss documents have also adopted a predominantly universalist approach. Most references to learners were broad, such as \u0026ldquo;at all school levels\u0026rdquo; and \u0026ldquo;regardless of origin, socioeconomic status, or physical and cognitive abilities.\u0026rdquo; Only a few more specific mentions appeared, such as references to \u0026ldquo;dyslexia and dyscalculia\u0026rdquo; and \u0026ldquo;particularly sensitive personal data\u0026rdquo; including \u0026ldquo;health data.\u0026rdquo; Similar to the Austrian documents, the Swiss documents did not provide detailed differentiation of individual groups, instead emphasizing inclusivity through general, all-encompassing language.\u003c/p\u003e\n\u003ch2\u003eSystemic Conditions\u0026nbsp;\u003c/h2\u003e\n\u003cp\u003eWhen analyzing the systemic condition, the German documents once again led the comparison to Austria and Switzerland, especially in terms of code frequency and their interpretation. The main focus across all countries in this theme was AI literacy. The importance of AI literacy was frequently emphasized in German documents (n = 37), which also referred to economic and access factors, stating it as a prerequisite for reducing the digital divide. One representative statement for this is: \u0026ldquo;A basic understanding of how the technology functions, as well as the ability to recognize the opportunities, limitations, and risks of new digital possibilities, is essential for a responsible and reflective use, as well as for promoting learning in the classroom and at home\u0026rdquo; (see document number 4, p. 6). Additionally, in several passages, concerns were expressed about equitable economic access: \u0026ldquo;The first-level divide described above must be addressed through appropriate measures. The federal states therefore aim to provide a joint interface solution that enables data-protection-compliant and cost-free access to LLMs in the school sector in the near future\u0026rdquo; (4, p. 14). However, the initial and ongoing training of teachers was discussed only once in the context of inclusion and equity.\u003c/p\u003e\n\u003cp\u003eIn Austria, AI literacy was mentioned less often (n = 5). Statements such as: \u0026ldquo;The goal is the responsible and pedagogically/didactically meaningful use of AI-based tools as support for teachers and learners, always with critical reflection on the framework conditions and possible consequences, such as data protection (personal data)\u0026rdquo; (1, p. 22) suggest a stronger emphasis on learning with AI rather than learning about AI. Neither access to devices nor the professional teacher development were explicitly addressed as a systemic condition. This suggests that the structural framework requirements remain largely underrepresented in Austrian documents.\u003c/p\u003e\n\u003cp\u003eIn the selected Swiss documents, AI literacy was also addressed less often than Germany (n = 6). One example stated: \u0026ldquo;Students must learn to engage competently and responsibly with this reality and its consequences (AI literacy). Children and adolescents should be supported in primary school on their way to a mature and informed use of generative ML functions\u0026rdquo; (10, p. 10). Additionally, there was one mention each of technical infrastructure and teacher training, noting: \u0026ldquo;Teacher educators at universities of teacher education continue their professional development and can demonstrate in both pre-service and in-service training how generative ML functions can be used for diagnostics and support planning\u0026rdquo; (10, p. 8).\u003c/p\u003e\n\u003ch2\u003eAI\u0026rsquo;s Exclusion Risks\u0026nbsp;\u003c/h2\u003e\n\u003cp\u003eWhen discussing the potential risks that AI may pose in education, the German AI policy papers identified numerous exclusion risks, and Germany was the only one among the three countries that addressed all subcategories of this theme, showing a broader awareness of exclusion issues. For example, one passage stated: \u0026ldquo;It is problematic that LLMs are not equally accessible to all learners. Barriers to access may arise from limited language proficiency, dyslexia, or visual impairments. Learners from low socio-economic backgrounds may face financial obstacles due to private providers and lack of support from parents\u0026rdquo; (5, p. 17). Bias and distortion (n = 8) were among the most frequently addressed issues, with specific focus on topics such as:\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026ldquo;A phenomenon closely related to hallucinations in the functioning of large language models (LLMs) is bias\u0026mdash;systematic distortions of information (Navigli et al., 2023), which arise from the selection of training data (Gombert et al., 2023). These biases could influence a wide range of dimensions, including moral orientation (Schramowski et al., 2022), religion, gender, ethnicity, profession (Nadeem et al., 2021), or political alignment (Feng et al., 2023; Rozado, 2023). Different LLMs may, for instance, adopt different political stances (Feng et al., 2023) or reproduce prejudices (e.g., Bergener et al., 2023)\u0026ldquo; (6, p. 9).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTechnological barriers were also discussed (n = 6), identifying the digital divide as a major challenge for educational equity. Cognitive barriers, including linguistic and cultural challenges, were mentioned in three segments, along with socio-economic factors. Discrimination and linguistic-cultural barriers were addressed only once.\u003c/p\u003e\n\u003cp\u003eComparably to Germany, the Austrian documents frequently addressed bias and distortions (n = 6) as well as discrimination (n = 5). An exemplary segment highlighting discrimination related to marginalized groups reads as follows: \u0026ldquo;At the same time, it is important to prevent discrimination, support disadvantaged groups, and ensure diversity\u0026rdquo; (1, p. 31). However, these perspectives often remain normative and less concrete regarding their implementation. Other barriers, such as technological (n = 3) and socioeconomic aspects (n = 1), were only marginally represented. Cognitive and linguistic-cultural barriers have not been addressed. This suggests that structural and infrastructural conditions for access received comparatively limited attention in the Austrian documents.\u003c/p\u003e\n\u003cp\u003eIn contrast to Germany and Austria, Swiss documents displayed a somewhat different emphasis regarding exclusion risks. Technological barriers (n = 6) were most frequently addressed, highlighting functional and infrastructural aspects. Bias and distortions (n = 4) were also prominently mentioned, for instance: \u0026ldquo;The risks of AI must always be considered (distortions, sources, etc.)\u0026rdquo; (8, p. 6). Socioeconomic aspects were moderately represented (n = 3), while discrimination, cognitive barriers, and linguistic-cultural barriers appeared only once. However, it is important to note that these latter three dimensions were often mentioned together in a single coded segment, potentially distorting individual frequency counts: \u0026ldquo;The schools guarantee that all children have access to the new opportunities with AI. Regardless of origin, socioeconomic status, and physical or cognitive abilities, all children should be able to participate equally\u0026rdquo; (8, p. 2).\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eAI\u0026rsquo;s Inclusive Potentials\u003c/h2\u003e\n\u003cp\u003eWhen discussing the potential risks of AI in education, it is essential to equally consider its potential benefits. The analysis of German documents revealed a clear emphasis on AI\u0026rsquo;s capability for differentiation, individualization, and adaptivity (n = 51), highlighting its use primarily as a tool to tailor learning experiences to individual learner needs, including technical accessibility aspects such as subtitles. For example, the Secretariat of the Standing Conference of Ministers of Education and Cultural Affairs explicitly states: \u0026ldquo;They also assume that adaptive and AI-supported learning materials can have a positive effect on the acquisition of basic competencies\u0026rdquo; (4, pp. 6\u0026ndash;7). Additionally, assessment-related potentials such as feedback provision, diagnostics, and learning analytics were prominently acknowledged (n = 20). Cooperative learning (n = 8) and managing heterogeneity (n = 3) received some attention, whereas gamification was only marginally referenced (n = 2). Notably, the social presence of AI bots or robots was entirely absent (n = 0), suggesting a predominantly functional-technical rather than a socially interactive framing of AI\u0026rsquo;s educational potential. In the Austrian documents, the primary focus is on differentiation, individualization, and adaptivity (n = 19), emphasizing AI\u0026apos;s potential for tailoring learning experiences, including considerations of technical accessibility such as subtitles. Additionally, assessment-related potentials, such as feedback provision, diagnostics, and learning analytics, were mentioned (n = 6). Addressing heterogeneity within classrooms received some attention (n = 4), reflecting an awareness of educational diversity and associated challenges. In contrast, the social presence of AI bots or robots was mentioned only marginally (n = 1). Cooperative learning and gamification potentials were entirely absent.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSwiss documents showed limited, yet focused, engagement with AI\u0026rsquo;s educational potentials. Differentiation, individualization, and adaptivity were mentioned three times (n = 3), indicating an understanding of AI as a means for personalized learning. assessment-related aspects such as feedback and diagnostics were also addressed (n = 2), alongside references to managing heterogeneity (n = 2), suggesting some recognition of diverse learner needs. However, no studies have been found regarding cooperative learning, gamification, or the social presence of AI bots or robots. This absence reinforces the impression that Swiss documents primarily frame AI as a technical and individualized support tool, with little emphasis on its interactive or socially embedded dimensions.\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eNormative and Participatory Frameworks\u003c/h2\u003e\n\u003cp\u003eWhile analyzing the policy documents, we also aimed to determine whether underlying inclusive frameworks were present. The analysis revealed that the AI policy documents from the DACH region neither explicitly address inclusion and equity nor reference international frameworks such as the UN Convention on the Rights of Persons with Disabilities (UN CRPD) or the Sustainable Development Goals (SDGs) related to inclusive education. Furthermore, only four passages were found that addressed participation and co-determination, understood as the involvement of students, persons with disabilities, or other affected groups in decision-making or design processes. Three of these references appear in the Secretariat of the Standing Conference of the Ministers of Education and Cultural Affairs policies in Germany (4). Among other aspects, they highlighted that informational self-determination is a fundamental right of children and adolescents that must be respected when using AI-based systems. One passage addressed the involvement of teachers and students in selecting AI applications. Another passage emphasized participation as an important educational goal. The fourth reference, appearing in the Austrian AIM document (1), stats that AI development should be human-centered and that participation should be a core design principle. Although the publisher, the Federal Ministry of Innovation, Mobility, and Infrastructure Republic Austria, states that over 160 experts from different fields (i.e. technology, economics, natural sciences, law, social sciences, and education) were involved in writing the paper, it\u0026rsquo;s unclear who participated or if they have expertise in inclusive education. However, none of the documents provided a detailed explanation or operationalization of the concept of participation.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe comparative analysis of the selected documents reveals a consistent gap between rhetorical commitments and policies for inclusion and equity in the AI field across the DACH countries. While Germany shows broader engagement, particularly in terms of diverse target group mentions, system-level conditions, and risk awareness, its policy documents still fall short of embedding inclusion as a binding governance principle. Austria and Switzerland demonstrate even less thematic depth, often relying on declarative, generalized statements without concrete mechanisms. These patterns reflect what Varsik and Vosberg (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) describe as the \u0026ldquo;superficial engagement\u0026rdquo; with inclusive digital transformation, highlighting the disconnect between AI's transformative potential and the ongoing marginalization of inclusion and equity as policy priorities.\u003c/p\u003e\u003cp\u003eAlthough Germany uses terms related to inclusion and equity more frequently than Austria and Switzerland do, their overall occurrence remains low. This suggests that inclusion and equity are only marginally addressed in the AI policy documents of all three DACH countries. Although Germany nominally leads with 12 mentions of inclusion and equity-related key terms, this picture becomes more nuanced when considering the significantly greater length and scope of its policy documents. This highlights a structural pattern of marginal visibility of inclusion and equity across the DACH region. It is important to note that these findings are based solely on frequency analysis, without assessing qualitative differences in thematic depth. For instance, the policy paper from the Standing Conference of the Ministers of Education and Cultural Affairs (KMK) in Germany dedicated an entire chapter to equity, illustrating a structural focus on the issue. Such a detailed treatment is absent in Austrian or Swiss documents. These findings should therefore be viewed as initial indicators of the theme\u0026rsquo;s visibility.\u003c/p\u003e\u003cp\u003eThe cross-country comparison reveals noticeable differences in how target groups are addressed in the AI policy papers of Germany, Austria, and Switzerland. Germany demonstrates the widest range of formulations, suggesting a more nuanced understanding of educational inequalities and a more targeted approach to vulnerable groups, although the language used can sometimes be vague or lack systematic operationalization. In contrast, the Austrian policy papers emphasize broader target groups. While there are references to \u0026ldquo;disadvantaged groups\u0026rdquo; and gender-specific support goals (\u0026ldquo;women and girls\u0026rdquo;), the overall language remains mainly declarative. The mention of \u0026ldquo;learners across the entire educational pathway\u0026rdquo; indicates a systemic understanding of education but lacks detailed reference to varying support needs. While all three DACH countries express general commitments to inclusion in their AI-in-education policies, only Germany shows tentative efforts toward a differentiated approach to target groups. Austria and especially Switzerland rely predominantly on universal formulations, often referring to the entire student population without specifying particular needs. This raises questions about the depth and effectiveness of their inclusion claims. Drawing on the work of Ainscow (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), inclusive education is understood not merely as a universal provision for all students but as a dynamic process that also requires a targeted emphasis on those groups at risk of marginalization, exclusion, or underachievement. Since inclusive education is founded on the belief that education is a fundamental human right and a cornerstone for a more just society, it is crucial to explicitly name and address the needs of vulnerable groups. Without such specificity, policies may overlook the unique challenges these groups face, thereby diluting the equity and fairness that inclusive policies aim to achieve. Following Ainscow\u0026rsquo;s (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) framework, inclusive education entails not only the physical presence of all learners but also their active participation and achievement. Most AI policies, however, primarily focus on presence, with little attention to systemic participation or tailored support mechanisms. This limited conceptualization also neglects the form of \u0026lsquo;representational participation\u0026rsquo; emphasized by Zowghi and da Rimini (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), where inclusion requires the active involvement and visibility of diverse learners in both the development and deployment of AI systems. From this perspective, the lack of specific references to vulnerable groups in Austrian and Swiss documents may reflect a gap between stated inclusive ideals and the operationalization of those ideals in policy. Naming and addressing specific groups, such as learners with disabilities, socioeconomically disadvantaged students, or those with learning differences, is essential for ensuring that equity is not merely aspirational but actionable within AI-driven educational transformations, as stated by Shams, Zowghi, and Bano (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eGermany addresses systemic conditions most comprehensively, especially AI literacy. While Switzerland and Austria also touch on this topic, their coverage is less extensive. Notably, Germany recognizes the importance of infrastructure and teacher training as relevant components for effective AI integration into education, whereas these aspects are nearly absent in Austrian documents. The near-complete absence of explicit teacher training programs in the Austrian and Swiss documents reveals structural capacity gaps in transferring AI literacy into pedagogical practice. The comparison reveals that systemic requirements are prioritized differently across the region, with Germany taking a more proactive stance. These disparities matter, as teacher training is widely regarded as essential for equitable adaptation of AI in education. Without an understanding of AI, there is a risk of uncritical use that may worsen education inequities (Mouta, Pinto-Liorente, \u0026amp; Torrecilla-S\u0026aacute;nchez, 2024). Continued training that addresses ethical and sociopolitical issues is crucial, enabling teachers to engage with AI in a critical and responsible manner (Jackson-Summers et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eGermany also demonstrated the most comprehensive engagement with exclusion risks. All seven subcodes were coded at least once, indicating a nuanced awareness of the issue. Austria takes a more selective approach, mainly focusing on discrimination and stereotypical bias, while neglecting technical access issues. While Austria focuses heavily on the debate around algorithmic bias, systemic dimensions, such as infrastructural inequalities, remain largely peripheral. In contrast, Switzerland concentrates primarily on technical barriers, with social and intersectional perspectives receiving little attention. Compared to Germany, the thematic range in Switzerland appears to be more limited, but it is clearly focused on learning-oriented adaptation. As Shams, Zowghi, and Bano (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2025\u003c/span\u003e) emphasize, AI systems that are developed without critical reflection on their data foundations and design logic risk not only replicating but reinforcing existing inequalities. In the analyzed policies, however, such structural forms of bias are rarely addressed explicitly or accompanied by concrete countermeasures. Switzerland finds a middle ground: various barriers are acknowledged, but with limited depth. Overall, the codings suggest that Swiss documents frame exclusion risks predominantly through a technological lens, with comparatively less focus on social or intersectional perspectives. As Shams, Zowghi, and Bano (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2025\u003c/span\u003e) argue, unexamined training data and implicit design assumptions can cause the systemic reproduction of educational inequalities. Representative justice, ensuring that diverse learners are visible and considered in AI development, is not meaningfully addressed in any of the reviewed documents. This absence reinforces the impression that Swiss documents primarily frame AI as a technical and individualized support tool, with little emphasis on its interactive or socially embedded dimensions. The complete omission of socially interactive AI scenarios, such as robotics applications designed to foster social-emotional competencies, highlights a narrow, instrumental-functional AI discourse that underestimates relational learning processes. This suggests significant divergence in national perspectives on inclusion-related challenges in AI: Germany adopts a more systemic approach, Austria emphasizes social justice in a narrower sense, and Switzerland mainly concentrates on linguistic and technological aspects.\u003c/p\u003e\u003cp\u003eThe analysis shows that Germany most frequently identified potential uses of AI, especially in support areas such as \u0026ldquo;differentiation, individualization/adaptivity,\u0026rdquo; which were coded 51 times. Austria explores similar potentials but less frequently. Notably, neither Switzerland nor Austria mentions cooperative learning, although it appears multiple times in German documents. Gamification remains a marginal topic across all three countries. Overall, these findings suggest that AI\u0026rsquo;s potential is mainly understood in functional terms, however, only Germany reflects a broader range of didactic opportunities particularly focusing on GenAI. Possibilities such as diagnostics with AI or augmented reality were not mentioned.\u003c/p\u003e\u003cp\u003eOnly Germany briefly mentioned participatory principles, including references to informational self-determination and stakeholder involvement. Austria and Switzerland remain entirely declarative or silent on this matter. Across all three countries, references to international frameworks such as the UN CRPD or SDGs are notably absent. The disregard for international reference frameworks undermines the stated ambition of aligning national AI policies with global goals for inclusion. Overall, participation and normative frameworks are weakly developed and mostly symbolic, rather than actionable. This aligns with Armstrong, Armstrong, and Spandagou\u0026rsquo;s (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2011\u003c/span\u003e) critique that inclusion often remains symbolic rather than structural, as reflected in the analyzed documents, which exhibit a rhetorical rather than operational commitment to participation and co-determination. From an \u0026lsquo;equity-by-design\u0026rsquo; perspective (WEF, 2022), effective inclusion demands not only generalized commitments but the deliberate implementation of design principles that reflect the needs and identities of historically marginalized groups. The absence of representational justice in these documents, manifested in the lack of explicit group references and inclusive governance mechanisms, limits the transformative potential of AI as outlined in UNESCO\u0026rsquo;s and WEF\u0026rsquo;s global frameworks (UNESCO, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; WEF, 2022).\u003c/p\u003e\u003cp\u003eGiven this context, the findings emphasize the need to integrate inclusion and participation into AI policy explicitly and structurally. As outlined in the theoretical framework, inclusion involves more than access; it requires representational justice, differentiated support, and participatory governance (UNESCO, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Armstrong et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). The near absence of references to international equity frameworks, such as the UN CRPD or SDGs, presents a missed opportunity to align national policies with global commitments to educational justice. Transitioning from symbolic gestures to binding actions is crucial to ensure that AI development genuinely promotes inclusive and equitable learning environments.\u003c/p\u003e\u003cp\u003e\u003cb\u003eConclusion and Implications for Policy, Education, Research\u003c/b\u003e\u003c/p\u003e\u003cp\u003eIn light of the preceding analysis, the following recommendations should be viewed not as supplementary measures but as an essential framework for fostering inclusive digital policy. The analysis of AI policy documents across the DACH countries reveals significant gaps in the integration of inclusion and equity principles into national AI strategic papers. Although Germany addresses AI literacy more extensively than Austria and Switzerland, this focus often remains detached from concrete, inclusive measures or differentiated implementation for marginalized groups. This lack of specificity poses challenges, as genuinely inclusive education requires explicitly targeting vulnerable populations and actively dismantling structural barriers (Ainscow, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Germany`s policies identify a broader range of exclusion risks and potentials of AI, including individualized and cooperative learning methods. In contrast, Austrian and Swiss documents predominantly frame AI\u0026rsquo;s potential through personalization and assessment, with limited acknowledgment of associated risks such as technological and socioeconomic barriers. The overarching emphasis on AI as a solution for individualized learning, particularly in German and Austrian documents, risks reinforcing a techno-solutionist narrative (Mochizuky et al., 2025) that overlooks structural inequalities and the need for accessible, participatory digital learning environments (Autenrieth et al., 2025). To avoid the \u0026ldquo;techno-ableism\u0026rdquo; trap, where AI tools are assumed to be neutral or universally effective, future policies or recommendations should be based on intersectional perspectives and co-designed with inclusion experts (Shew, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). As emphasized by Autenrieth et al. (2025), transformative AI education requires more than just technological infrastructure; it needs ethical reflection, normative considerations, and participatory governance to promote educational justice. Before stating the recommendations of this paper, it should be noted that all three countries notably omit international equity frameworks, such as the UN CRPD and the SDGs, thereby missing an opportunity to align with global commitments to educational justice (UNESCO, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2017\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eFor future policy and research, several recommendations have emerged from the results. First, inclusion must be established as a binding principle within national AI policies guided explicitly by international frameworks (UN-CRPD, SDGs). Second, systematic identification and targeted support for vulnerable learner groups should become standard practice. Third, inclusive AI competencies should be embedded comprehensively in teacher education to foster critical, equitable AI use (Autenrieth et al., 2025). Fourth, while the German documents provide isolated examples of good practices, they lack systematic evaluation. A research agenda focusing on the evaluation of these within real school settings should be supported by national policy funding. Finally, participatory governance mechanisms involving different stakeholders, such as educators, inclusion experts, and affected communities should be strengthened (Shams, Zowghi, \u0026amp; Bano, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). This approach demands a shift from techno-centric to democratic and inclusive AI governance models, which is crucial for realizing the transformative potential of AI in education. In line with UNESCO\u0026rsquo;s (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2005\u003c/span\u003e) call for a systemic transformation of educational systems and Autenrieth et al.\u0026rsquo;s (2025) vision of a \u0026lsquo;democratic logic\u0026rsquo; in AI education, future policy frameworks must be grounded in inclusive governance, equity-by-design, and representational justice.\u003c/p\u003e\u003cp\u003eTo conclude, the document analysis revealed that inclusion and equity principles are not yet firmly embedded in AI educational policy papers. To move beyond symbolic rhetoric, it is essential to openly address participation, non-discrimination, and diversity, integrating these principles into both educational design and regulatory frameworks (UNICEF, 2017). This aligns with Autenrieth et al.`s (2025) call for a shift from an \u0026ldquo;innovation logic\u0026rdquo; to a \u0026ldquo;democratic logic\u0026rdquo; in AI governance in education. Achieving truly inclusive AI education requires more than rhetorical alignment; it demands a comprehensive shift from innovation-centric models to participatory, justice-based policies and strategies.\u003c/p\u003e\u003cp\u003e\u003cb\u003eLimitations of the Study\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThis study has several limitations. First, it focuses only on official policies, excluding non-official documents such as reports from NGOs and educational institutions. This means that it does not capture how the policies are implemented in practice or their impact on educational equity and inclusion. Second, the analysis includes documents published between 2021 and May 2025, which may miss recent updates or drafts, especially those related to developments around the EU AI Act. Third, the framework used for inclusion and equity reflects dominant Western perspectives, and the qualitative coding process involves some subjectivity, despite efforts to mitigate this through independent coding and consensus discussions. Future research could improve this by triangulating the findings with interviews or focus groups with policymakers or other stakeholders to validate document interpretations. Fourth, this study\u0026rsquo;s geographical and cultural context is also a limitation. While the DACH region shares some cultural and political similarities, there are important differences that could affect how AI and inclusion are implemented and understood in education. Future research might include examining sub-national differences within the DACH region.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eFunding\u003c/h2\u003e\n\u003cp\u003eThe authors declare that no funds, grants, or other support were received during the preparation of this manuscript.\u003c/p\u003e\n\u003ch2\u003eCompeting Interests\u003c/h2\u003e\n\u003cp\u003eThe authors have no relevant financial or non-financial interests to disclose.\u003c/p\u003e\n\u003ch2\u003eAuthor Contributions\u003c/h2\u003e\n\u003cp\u003eAll authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by Edvina Be\u0026scaron;ić, Christa Schmid-Meier, and Lea Schulz. The first draft of the manuscript\u0026apos;s introduction was written by Christ Schmid-Meier, the Methods section by Edvina Be\u0026scaron;ić, and the analysis by Lea Schulz and Edvina Be\u0026scaron;ić. The other parts of the paper were written in collaboration by the three authors, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAhmed, S., Rahman, M. S., Kaiser, M. S., \u0026amp; Hosen, A. S. M. S. (2025). Advancing Personalized and Inclusive Education for Students with Disability Through Artificial Intelligence: Perspectives, Challenges, and Opportunities. \u003cem\u003eDigital\u003c/em\u003e, \u003cem\u003e5\u003c/em\u003e(2), 11. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/digital5020011\u003c/span\u003e\u003cspan address=\"10.3390/digital5020011\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAinscow, M. (2020). Promoting inclusion and equity in education: lessons from international experiences. \u003cem\u003eNordic Journal of Studies in Educational Policy\u003c/em\u003e, \u003cem\u003e6\u003c/em\u003e(1), 7\u0026ndash;16. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/20020317.2020.1729587\u003c/span\u003e\u003cspan address=\"10.1080/20020317.2020.1729587\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eArmstrong, D., Armstrong, A. C., \u0026amp; Spandagou, I. (2011). Inclusion: By choice or by chance? \u003cem\u003eInternational Journal of Inclusive Education\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(1), 29\u0026ndash;39. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/13603116.2010.496192\u003c/span\u003e\u003cspan address=\"10.1080/13603116.2010.496192\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAutenrieth, D., Schluchter, J. R., \u0026amp; Schulz, L. (in press). AI is all you need? Zeitschrift f\u0026uuml;r Inklusion.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eArias-Flores, H., Valencia-Arag\u0026oacute;n, K., Calle-Jimenez, T., \u0026amp; Sanchez-Gordon, S. (2025). Artificial Intelligence and Assistive Technologies: A Systematic Review of Educational Applications for Disabilities. In M. Antona \u0026amp; C. Stephanidis (Eds.), \u003cem\u003eUniversal Access in Human-Computer Interaction\u003c/em\u003e (Vol. 15780, pp. 283\u0026ndash;292). Springer Nature Switzerland. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-3-031-93848-1_19\u003c/span\u003e\u003cspan address=\"10.1007/978-3-031-93848-1_19\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDillenbourg, P. (2016). The Evolution of Research on Digital Education. \u003cem\u003eInternational Journal of Artificial Intelligence in Education\u003c/em\u003e, \u003cem\u003e26\u003c/em\u003e(2), 544\u0026ndash;560. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s40593-016-0106-z\u003c/span\u003e\u003cspan address=\"10.1007/s40593-016-0106-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDong, M., Conway, J. R., Bonnefon, J. F., Shariff, A., \u0026amp; Rahwan, I. (2024). Fears about artificial intelligence across 20 countries and six domains of application. \u003cem\u003eAmerican Psychologist\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1037/amp0001454\u003c/span\u003e\u003cspan address=\"10.1037/amp0001454\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDua, M., Singh, J. P., \u0026amp; Shehu, A. (2025). The ethics of national artificial intelligence plans: an empirical lens. \u003cem\u003eAI and Ethics\u003c/em\u003e, \u003cem\u003e5\u003c/em\u003e(4). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s43681-025-00663-2\u003c/span\u003e\u003cspan address=\"10.1007/s43681-025-00663-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFederal Office of Communications (2025). Artificial Intelligence: Overview and Switzerland\u0026rsquo;s regulatory approach. \u003cem\u003eArtificial Intelligence\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.bakom.admin.ch/bakom/en/homepage/digital-switzerland-and-internet/strategie-digitale-schweiz/ai.html\u003c/span\u003e\u003cspan address=\"https://www.bakom.admin.ch/bakom/en/homepage/digital-switzerland-and-internet/strategie-digitale-schweiz/ai.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGhimire, A., \u0026amp; Edwards, J. (2024). From Guidelines to Governance: A Study of AI Policies in Education. In A. M. Olney, I.-A. Chounta, Z. Liu, O. C. Santos, \u0026amp; I. I. Bittencourt (Eds.), \u003cem\u003eArtificial Intelligence in Education. Posters and Late Breaking Results, Workshops and Tutorials, Industry and Innovation Tracks, Practitioners, Doctoral Consortium and Blue Sky\u003c/em\u003e (Vol. 2151, pp. 299\u0026ndash;307). Springer Nature Switzerland. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-3-031-64312-5_36\u003c/span\u003e\u003cspan address=\"10.1007/978-3-031-64312-5_36\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJackson-Summers, A. G., Mrakovcich, K. L., Gray, J. P., Fleischmann, C. M., Emami, T., \u0026amp; Page, E. J. (2024). A systematic review of inclusive pedagogical research using the CIRTL inclusive pedagogy framework: multi-disciplinary and STEM perspectives, current trends and a research agenda. \u003cem\u003eDiscover Education\u003c/em\u003e, \u003cem\u003e3\u003c/em\u003e(1), 30. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s44217-024-00093-y\u003c/span\u003e\u003cspan address=\"10.1007/s44217-024-00093-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKent, C., \u0026amp; Boulay, B. D. (2022). \u003cem\u003eAI for Learning\u003c/em\u003e (1st ed.). CRC. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1201/9781003194545\u003c/span\u003e\u003cspan address=\"10.1201/9781003194545\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKozleski, E., Artiles, A., \u0026amp; Waitoller, F. (2014). Equity in inclusive education: A cultural historical comparative perspective. In \u003cem\u003eThe SAGE handbook of special education\u003c/em\u003e (Vol. 2, pp. 231\u0026ndash;245). SAGE. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4135/9781446282236.n16\u003c/span\u003e\u003cspan address=\"10.4135/9781446282236.n16\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKuckartz, U., \u0026amp; R\u0026auml;diker, S. (2024). \u003cem\u003eQualitative Inhaltsanalyse. Methoden, Praxis, Umsetzung mit Software und k\u0026uuml;nstlicher Intelligenz\u003c/em\u003e (6. Auflage) [Qualitative Content Analysis: Methods, Practice, Implementation with Software and Artificial Intelligence]. Juventa Verlag.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMelo-L\u0026oacute;pez, V. A., Basantes-Andrade, A., Gudi\u0026ntilde;o-Mej\u0026iacute;a, C. B., \u0026amp; Hern\u0026aacute;ndez-Mart\u0026iacute;nez, E. (2025). The Impact of Artificial Intelligence on Inclusive Education: A Systematic Review. \u003cem\u003eEducation Sciences\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(5), 539. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/educsci15050539\u003c/span\u003e\u003cspan address=\"10.3390/educsci15050539\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMochizuki, Y., Bruillard, E., \u0026amp; Bryan, A. (2025). The ethics of AI or techno-solutionism? UNESCO\u0026rsquo;s policy guidance on AI in education. \u003cem\u003eBritish Journal of Sociology of Education\u003c/em\u003e, 1\u0026ndash;22. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/01425692.2025.2502808\u003c/span\u003e\u003cspan address=\"10.1080/01425692.2025.2502808\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMonett, D., \u0026amp; Paquet, G. (2025). Against the Commodification of Education \u0026mdash; if harms then not AI. \u003cem\u003eJournal of Open Distance and Digital Education\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.25619/WAZGW457\u003c/span\u003e\u003cspan address=\"10.25619/WAZGW457\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMouta, A., Pinto-Llorente, A. M., \u0026amp; Torrecilla-S\u0026aacute;nchez, E. M. (2024). Uncovering Blind Spots in Education Ethics: Insights from a Systematic Literature Review on Artificial Intelligence in Education. \u003cem\u003eInternational Journal of Artificial Intelligence in Education\u003c/em\u003e, \u003cem\u003e34\u003c/em\u003e(3), 1166\u0026ndash;1205. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s40593-023-00384-9\u003c/span\u003e\u003cspan address=\"10.1007/s40593-023-00384-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eOECD (2024). \u003cem\u003eExplanatory memorandum on the updated OECD definition of an AI system\u003c/em\u003e (OECD Artificial Intelligence Papers No. 8). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1787/623da898-en\u003c/span\u003e\u003cspan address=\"10.1787/623da898-en\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eOECD (2025). \u003cem\u003eEmpowering learners for the age of AI: An AI literacy framework for primary and secondary education (Review draft)\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://ailiteracyframework.org/?utm_source=substack\u0026amp;utm_medium=email\u003c/span\u003e\u003cspan address=\"https://ailiteracyframework.org/?utm_source=substack\u0026amp;utm_medium=email\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSalas-Pilco, S. Z. (2020). The impact of AI and robotics on physical, social‐emotional and intellectual learning outcomes: An integrated analytical framework. \u003cem\u003eBritish Journal of Educational Technology\u003c/em\u003e, \u003cem\u003e51\u003c/em\u003e(5), 1808\u0026ndash;1825. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/bjet.12984\u003c/span\u003e\u003cspan address=\"10.1111/bjet.12984\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSchiff, D. (2022). Education for AI, not AI for Education: The Role of Education and Ethics in National AI Policy Strategies. \u003cem\u003eInternational Journal of Artificial Intelligence in Education\u003c/em\u003e, \u003cem\u003e32\u003c/em\u003e(3), 527\u0026ndash;563. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s40593-021-00270-2\u003c/span\u003e\u003cspan address=\"10.1007/s40593-021-00270-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSchmid-Meier, C., \u0026amp; Schulz, L. (2025). Exploring the Inclusive Potential of AI Chatbots: Engagement and Barriers in Foreign Language Training. In M. Antona \u0026amp; C. Stephanidis (Eds.), \u003cem\u003eUniversal Access in Human-Computer Interaction\u003c/em\u003e (Vol. 15780, pp. 377\u0026ndash;395). Springer Nature Switzerland. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-3-031-93848-1_25\u003c/span\u003e\u003cspan address=\"10.1007/978-3-031-93848-1_25\" 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. (2025). AI and the quest for diversity and inclusion: a systematic literature review. \u003cem\u003eAI and Ethics\u003c/em\u003e, \u003cem\u003e5\u003c/em\u003e(1), 411\u0026ndash;438. \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\u003eShew, A. Ableism, Technoableism, and, \u0026amp; Future, A. I. (2020). \u003cem\u003eIEEE Technology and Society Magazine\u003c/em\u003e, \u003cem\u003e39\u003c/em\u003e(1), 40\u0026ndash;85. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/mts.2020.2967492\u003c/span\u003e\u003cspan address=\"10.1109/mts.2020.2967492\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eUNESCO (2005). \u003cem\u003eGuidelines for Inclusion: Ensuring Access to Education for All. UNESCO.\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eUNESCO. (2017). \u003cem\u003eA Guide for ensuring inclusion and equity in education\u003c/em\u003e. UNESCO.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003evan Dijk, J. (2020). \u003cem\u003eThe Digital Divide\u003c/em\u003e. Polity.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVarsik, S., \u0026amp; Vosberg, L. (2024). \u003cem\u003eThe potential impact of Artificial Intelligence on equity and inclusion in education\u003c/em\u003e (OECD Artificial Intelligence Papers No. 23). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1787/15df715b-en\u003c/span\u003e\u003cspan address=\"10.1787/15df715b-en\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWahono, Z. (2025). Integrating AI into Constructivist Pedagogy: Strategies and Outcomes. \u003cem\u003eAdvances in Nonlinear Variational Inequalities\u003c/em\u003e, \u003cem\u003e28\u003c/em\u003e(5s), 227\u0026ndash;231. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.52783/anvi.v28.3789\u003c/span\u003e\u003cspan address=\"10.52783/anvi.v28.3789\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZowghi, D., \u0026amp; da Rimini, F. (2023). Diversity and Inclusion in Artificial Intelligence. arXiv preprint arXiv:2305.12728, 2023.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWorld Economic Forum [WEF] (2022). \u003cem\u003eA Blueprint for Equity and Inclusion in Artificial Intelligence\u003c/em\u003e. (n.d.). Retrieved July 19, 2025, from \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www3.weforum.org/docs/WEF_A_Blueprint_for_Equity_and_Inclusion_in_Artificial_Intelligence_2022.pdf\u003c/span\u003e\u003cspan address=\"https://www3.weforum.org/docs/WEF_A_Blueprint_for_Equity_and_Inclusion_in_Artificial_Intelligence_2022.pdf\" targettype=\"URL\" 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":false,"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":"Artificial Intelligence Policies, Inclusive Education, Educational Equity, DACH region","lastPublishedDoi":"10.21203/rs.3.rs-7232992/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7232992/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis paper examines how inclusion and educational equity are conceptualized and thematized in national AI and education policies in Germany, Austria, and Switzerland (the DACH region). Through a qualitative content analysis of 10 policy documents, this study explores how these policies articulate inclusion and equity, and whether they include concrete implementation plans. The analysis was conducted using MAXQDA, combining inductive codes related to inclusion and equity with a deductive framework. The policy analysis reveals a contrast in how the DACH region approaches inclusive and equitable AI education. Germany uses explicit equity metrics and funding, Austria relies on implicit STEM-focused inclusion, and Switzerland\u0026rsquo;s decentralized model yields fragmented, declarative measures, all three lacking enforceable commitments or measurable targets. This paper presents the first comparative analysis of how inclusion and equity are addressed in AI policy documents in the DACH region. It highlights promising examples of inclusive governance while also pinpointing structural barriers that may impede equitable access to AI-supported learning. Building on these insights, this paper develops targeted recommendations for future policy development, advocating not only for the integration of inclusion and equity as core principles in national AI policies but also for conceiving inclusion as a collaborative, cross-border effort that leverages shared good practices and joint oversight.\u003c/p\u003e","manuscriptTitle":"Inclusion and Equity in Artificial Intelligence (AI): Analysis of Educational Policies in Austria, Germany, and Switzerland","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-15 18:09:07","doi":"10.21203/rs.3.rs-7232992/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":"25884c1b-bc2d-49c8-a58b-b83f67d808bc","owner":[],"postedDate":"October 15th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-03-19T08:41:52+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-15 18:09:07","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7232992","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7232992","identity":"rs-7232992","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}