A Systematic Comparative Review of Generative AI vs. Teacher Feedback in EFL Writing: Learner Perceptions, Preferences, and Psychological Factors

A Systematic Comparative Review of Generative AI vs. Teacher Feedback in EFL Writing: Learner Perceptions, Preferences, and Psychological Factors | 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 Systematic Review A Systematic Comparative Review of Generative AI vs. Teacher Feedback in EFL Writing: Learner Perceptions, Preferences, and Psychological Factors Yanmei Zhao¹, Dandan Li² This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8386846/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 13 You are reading this latest preprint version Abstract The introduction of GenAI into instruction in a second language puts the conventional focus on teacher feedback in question. Although tools such as ChatGPT are quickly spreading, whether algorithmic or human feedback are more effective in education is an issue in pedagogical discussion. The systematic review consolidates findings on 17 empirical studies (2023–2025) with a unique focus on GenAI and teacher feedback comparison in EFL writing situation. With reference to the self-determination and cognitive load theories, our analysis reveals the steady division of labor in feedback efficacy where GenAI performs more effectively as a nurturing factor in reducing anxiety and surface-level error-correction skills than teacher feedback, which performs better in fostering higher-order level argumentation, coherent and metacognitive awareness. Most importantly, this review empirically confirms the superiority in hybrid feedback models, which allow using AI to achieve immediacy and accuracy, with a human-focused viewpoint used only to respond to the situation. We suggest that GenAI should exist and operate in a culturally competent, collaborative structure intelligence that does not replace, but adds the scaffold that will streamline the work of the teacher and increase student autonomy. The results can serve as an evidence-based guideline to applying AI in writing classes and reducing the risks of overdependence. Physical sciences/Mathematics and computing Biological sciences/Psychology Social science/Psychology Generative AI EFL writing feedback Hybrid models Learner psychology Figures Figure 1 Introduction Written corrective feedback (WCF) has been one of the foundations of second language (L2) writing education, which is a critical factor in achieving linguistic accuracy and proficiency writing (Bitchener and Ferris, 2012; Hyland and Hyland, 2006). This is especially important in English as a Foreign Language (EFL) settings, when the availability of high-quality feedback is also usually very restricted. Although conventional teacher feedback is being cherished due to the ability to identify the higher-order issues and to be sensitive to context-based concerns (Ferris, 2006; Hyland, 2013), it is often undermined by the practical factors, including a high number of students in a classroom and the workload of a teacher (Zhang, 2020). This landscape can be changed with the emergence of generative artificial intelligence (AI), namely large language models such as ChatGPT. These systems offer instant, scalable, and linguistically advanced feedback, which may result in human-provided WCF overcoming some of its essential constraints (Barrot, 2023; Warschauer et al., 2023; Mekheimer, 2025). Participating in contrast to the previous automated writing assessors that tried to detect the errors only on the surface (Cobb, 2008), the modern generative AI can provide detailed feedback on the content, structure, and the rhetoric, and in some cases, it can even excel at human tasks. These potentials are highlighted by research on emerging comparative studies, which determine that GenAI is effective in alleviating writing anxiety and boosting lexical resources (Wang, 2024; Ramazani et al., 2025). Nevertheless, one of the constant conclusions of the recent literature is that it is relatively weak in building strong cognitive involvement, effective acquisition, and metacognitive awareness relative to the level of fine guidance of instructors (Muñoz Muñoz et al., 2025; Yu and Xie, 2025). Nevertheless, even with this quickly growing body of empirical research since 2023 the subject matter remains counterintuitive, conclusively, inconsistent. There is also no meta-analysis to date (Wang and Dang, 2024), the findings on qualitative level are mostly cultural siloed (Yang and Yao, 2025; ), and the most essential aspects such as perceptions, preferences, and psychological reactions of learners (e.g., a sense of autonomy, relational trust, writing anxiety) that ultimately mediate feedback uptake, have never been synthesised (Henderson et al., 2025; Shen et al., 2025). Moreover, even though hybrid human-AI feedback models enjoy a growing body of support (Zhang et al., 2025), they have minimal empirical basis and lack contextual differentiation (Derakhshan, 2025; Lo et al., 2025). There is therefore an urgent need to have a detailed systematic synthesis to justify these opposing results, consider contextual variables, and evaluate the empirical rationale upon the integration of the two information sources of feedback. This systematic review intends to bring together the still-expanding fragmented data on comparative research in the field of generative AI and teacher feedback in EFL writing guided by the self-determination theory (Ryan and Deci, 2020) and the cognitive load theory. It aims at fulfilling the following purposes: firstly, to critically scrutinize and resolve, the discrepant empirical evidence on their varied effects on the development of L2 writing, especially on complexity, accuracy, fluency, and overall quality; second, to unite qualitative and quantitative results concerning how learners perceive, how they like, and how they are psychologically inclined to respond to feedback, and third, to weigh up the empirical evidence upon emergent hybrid feedback models and to suggest a coherent and culturally aware framework through which they should be taught. Methodology Information Sources and Search Strategy A systematic literature search was performed across several major electronic databases, including Google Scholar, Web of Science, ScienceDirect, SpringerLink, and SAGE Journals. To capture the entire body of relevant literature since the emergence of generative AI, the search covered all publications from the inception of each database up to the date the search was conducted (November 2025). The search strategy was designed to be broad and inclusive, applying the following Boolean query to the “All Fields” metadata where possible: ( (EFL OR "English as a Foreign Language" OR "English learner" OR "L2 writer" OR "second language writ" OR "non-native speaker") ) AND ( (“Generative AI” OR “Generative Artificial Intelligence” OR “ChatGPT” OR “GPT-3” OR “GPT-4” OR “large language model” OR “LLM” OR “AI-generated feedback” OR “AI writing tool”) ) AND ( (“teacher feedback” OR “instructor feedback” OR “human feedback”) ). Eligibility Criteria Studies were selected based on the pre-defined PICOS framework: Population (P): Secondary or tertiary-level EFL learners. Intervention (I): Writing feedback generated by generative AI models (e.g., ChatGPT). Comparator (C): Feedback provided by teachers or human instructors. Outcomes (O): Comparative measures of writing performance (e.g., CAF, rubric scores), feedback uptake, or learner perceptions. Study Design (S): Empirical comparative studies (e.g., RCTs, quasi-experimental, mixed-methods). Exclusion criteria were: Non-empirical publications (e.g., reviews, theoretical papers). Studies without a direct comparison between generative AI and teacher feedback. Articles using non-generative AI systems (e.g., traditional AWE). Non-English publications. Study Selection Process The selection of the study adhered to PRISMA flow diagram (Fig. 1 ). A total of 2,623 records were initial; 500 records were duplicates. Titles and abstracts of 2123 unique records were screened with 1910 being excluded. The rest of the 213 articles were evaluated and 196 articles were eliminated because of certain reasons: lack of direct comparison (n = 121), the employment of non-generative AI (n = 48), and lack of data (n = 26). An ultimate sample size of 17 studies was added. This was done by two reviewers working individually and any difference was overcome by consensus. Data Extraction A systematic data extraction process was done in order to systematically identify the most important characteristics of the included studies, pointing towards a pre-defined checklist in accordance with the PRISMA guidelines (Page et al., 2021). Two reviewers (Y. Zhao and D. Li) independently identified data on four core domains ( 1 ) basic study information (authors, year, country, design, duration); ( 2 ) participant profile (sample size, level of proficiency); ( 3 ) intervention (AI tool and prompt design, teacher feedback characteristics, writing task); ( 4 ) key findings of the AI-teacher comparison. Minor differences among the reviewers were addressed by consensus discussion, and when the facing out of an issue was necessary, a third senior researcher was consulted. Quality Assessment All the studies (n = 18) included were evaluated with the Mixed Methods Appraisal Tool (MMAT) 2018 (Crowe et al., 2011; Pluye and Hong, 2014), which is a well-validated evaluation instrument of the quality of study designs (mixed methods, quantitative descriptive, non-randomized, RCT, and qualitative studies). The MMAT test included two obligatory screening questions (S1-S2) and criteria that depended on the research design (1.1–5.5). All three researchers had to train on MMAT 2018 before the formal assessment to ascertain their similarity in the definition of criteria and standards on how to pass marks. The two scholars (Y. Zhao and D. Li) independently rated each study. Poor ratings in terms of quality were sorted out through group discussion. The results of quality assessment showed that the quality of methodological approaches of all the considered studies was satisfactory. In particular, 15 articles (83.3% of all) were considered to be of high quality, these articles met all possible MMAT criteria. The rest of the studies (16.7% of all) were classified as having medium quality; they did not fail to satisfy the two screening questions or the majority of the design-specific criteria, but they were somehow limited (e.g., not a randomized study, not well integrated quantitative and qualitative findings in mixed methods studies). None of the studies were considered to be of low quality, which ensured that the used research materials were credible to carry out further analysis (See Table 1 ). Data Synthesis Since the extent of heterogeneity of experimental designs and outcomes measures of the studies included was very high, a narrative synthesis methodology was used in analyzing the data. Quantitative findings were summarized and shown descriptively and qualitative findings analyzed using a thematic analysis. This was sought to determine convergent and divergent trends of the comparative effectiveness of AI and teacher feedback, learners perceptions and preferences. Table 1 Summary of methodological quality assessment First Author, Year SCREENING QUESTIONS 1. QUALITATIVE STUDIES 2. RANDOMIZED CONTROLLED TRIALS 3. NON-RANDOMIZED STUDIES 4. QUANTITATIVE DESCRIPTIVE STUDIES 5. MIXED METHODS STUDIES Quality Rating S1 S2 1.1 1.2 1.3 1.4 1.5 2.1 2.2 2.3 2.4 2.5 3.1 3.2 3.3 3.4 3.5 4.1 4.2 4.3 4.4 4.5 5.1 5.2 5.3 5.4 5.5 Ramazani et al., 2025 ✔ ✔ - - - - - - - - - - - - - - - - - - - - ✔ ✔ ✔ ✔ ✔ high Solovey, 2024 ✔ ✔ ✔ ✔ ✔ - - - - - - - - - - - - - - - - - - - - - - high Cao et al., 2024 ✔ ✔ - - - - - - - - - - ✔ ✔ ✔ ✔ ✔ - - - - - - - - - - high Soori et al., 2025 ✔ ✔ - - - - - - - - - - ✔ ✔ ✔ × ✔ - - - - - - - - - - medium Wang, 2024 ✔ ✔ - - - - - - - - - - - - - - - - - - - - ✔ ✔ ✔ ✔ ✔ high Zhang et al., 2025 ✔ ✔ - - - - - - - - - - × ✔ ✔ × ✔ - - - - - - - - - - medium Muñoz et al.,2025 ✔ ✔ - - - - - - - - - - ✔ ✔ ✔ ✔ ✔ - - - - - - - - - - high Choi, 2024 ✔ ✔ - - - - - - - - - - - - - - - - - - - - ✔ ✔ ✔ ✔ ✔ high Abdi et al., 2025 ✔ ✔ - - - - - - - - - - ✔ ✔ ✔ ✔ ✔ - - - - - - - - - - high Yu et al., 2025 ✔ ✔ - - - - - - - - - - - - - - - - - - - - ✔ ✔ ✔ ✔ ✔ high Topuz et al., 2025 ✔ ✔ - - - - - - - - - - - - - - - ✔ ✔ ✔ ✔ ✔ - - - - - high Alnemrat et al., 2025 ✔ ✔ - - - - - - - - - - - - - - - ✔ ✔ ✔ ✔ ✔ - - - - - high Zou et al., 2025 ✔ ✔ - - - - - - - - - - - - - - - - - - - - ✔ ✔ ✔ ✔ ✔ high Makwana, 2025 ✔ ✔ - - - - - - - - - - - - - - - - - - - - ✔ ✔ ✔ ✔ ✔ high Aljasser, 2025 ✔ ✔ - - - - - - - - - - - - - - - - - - - - ✔ ✔ ✔ ✔ ✔ high Escalante et al., 2023 ✔ ✔ - - - - - - - - - - - - - - - - - - - - ✔ ✔ ✔ ✔ ✔ high Asadi et al., 2025 ✔ ✔ - - - - - - - - - - - - - - - - - - - - ✔ ✔ ✔ ✔ ✔ high N otes: S1 = Are there clear research questions?; S2 = Do the collected data allow to address the research questions?; 1.1 = Is the qualitative approach appropriate?; 1.2 = Are qualitative data collection methods adequate?; 1.3 = Are findings derived from data?; 1.4 = Is interpretation substantiated by data?; 1.5 = Is there coherence between components?; 2.1 = Is randomization appropriately performed?; 2.2 = Are groups comparable at baseline?; 2.3 = Are there complete outcome data?; 2.4 = Are outcome assessors blinded?; 2.5 = Did participants adhere to intervention?; 3.1 = Are participants representative?; 3.2 = Are measurements appropriate?; 3.3 = Are there complete outcome data?; 3.4 = Are confounders accounted for?; 3.5 = Was intervention administered as intended?; 4.1 = Is sampling strategy relevant?; 4.2 = Is sample representative?; 4.3 = Are measurements appropriate?; 4.4 = Is risk of nonresponse bias low?; 4.5 = Is statistical analysis appropriate?; 5.1 = Is there adequate rationale for mixed methods?; 5.2 = Are components effectively integrated?; 5.3 = Are integrated outputs adequately interpreted?; 5.4 = Are divergences between results addressed?; 5.5 = Do components adhere to quality criteria of each method?; ✔=Yes; ×=NO; “-” indicates that the criterion is not applicable to the study design. Results Characteristics of the Included Studies The scope of geographical distribution of the 17 analyzed studies is quite broad: the studies were carried out in East Asian (China, Hong Kong, South Korea) and Middle Eastern (Iran, Israel, Jordan, Saudi Arabia, Turkey) settings as well as in South Asian (India) and American (USA, Chile) settings. Such a variety highlights the international concern about the field of research (See Table 2 ). The studies had a variety of methods used, such as randomized controlled trials (RCTs), quasi-experimental, mixed-methods, and qualitative in the methodological aspect. The length of the interventions was not exactly the same, as short, intensive investigations of a week length were also interspersed with longitudinal studies lasting an entire semester of academia. The participants were students with A2 proficiency to postgraduate students (C2 proficiency, most of them at the undergraduate level, i.e. B1-C1). ChatGPT (with versions GPT-3.5, GPT-4, and GPT-4 Turbo) was the most commonly used AI tool, and prompt design became one of the key aspects that determine the quality of a piece of AI-generated feedback. Teacher feedback was also described as individualized, context-dependent and usually on the higher-order issues in writing. Table 2 Characteristics of Studies Included Authors (Year) Country Design; Duration Participants; Proficiency Writing Task AI Tool; Prompt Design Teacher Feedback Characteristics Core Comparison Dimensions Key Findings Ramazani et al. (2025) Iran Mixed-methods; 6 weeks 100 IELTS candidates; B2-level IELTS Writing Tasks ChatGPT 3.0; IELTS-aligned prompts Personalized annotations + consultations IELTS 4 criteria + learner perceptions Teacher better for higher-order; AI for surface errors; Hybrid recommended Solovey (2024) Israel Qualitative; NR 30 EFL students; B2 level Argumentative essay ChatGPT; Checklist prompts Written corrective feedback Perceived effectiveness; feedback preferences Teacher preferred for expertise/reliability; AI valued for speed/comprehensiveness. Cao et al. (2024 China Experimenta; 2-week intervals 45 MTI students; TEM-8 Chinese-English translation ChatGPT-4; standardized prompt Contextual feedback; translation logic; syntactic accuracy Translation quality; linguistic features (lexicon/syntax/cohesion) AI excels in lexical capacity; Teacher superior for syntax and overall quality; Blended approach recommended. Soori et al. (2025) Iran Randomized controlled tria; 2.5 months 88 IELTS EFL learners; Intermediate 200 word IELTS Task 2 essays ChatGPT-4, Grammarly; prompts: IELTS-aligned E-feedback; personalize; context-rich; higher-order skills IELTS criteria; writing performance Hybrid > AI or Teacher; AI > Teacher (vocab); Teacher > AI (higher-order); Students prefer hybrid. Wang (2024) China Quasi-experimental; 14 weeks 75 undergraduate EFL learners; NR Weekly course assignments Poe app; prompts: grammar/coherence Traditional corrective Anxiety; CAF AI > Teacher (anxiety, accuracy, fluency, complexity). Zhang et al. (2025) China Experimental; 12 weeks 60 Chinese EFL students; TOEFL > 90, IELTS > 7 GRE Issue Writing Task GPT-4 Turbo; GRE rubrics Personalized; supplements AI feedback; higher-order skills GRE writing criteria; motivation; feedback perception Hybrid feedback superior to AI-only; AI for basic skills but limited in higher-order thinking. Muñoz et al. (2025) Chile Quasi-experimental; Over 2 months 44 EFL students; CEFR A2–B1 IELTS-style opinion essays GPT-4o; Role-playing as English teacher Direct written corrective; Band scores IELTS criteria ChatGPT > Teacher (all criteria); Large effect size for AI. Choi (2024) South Korea Mixed-methods; NR 20 high school students; CEFR B2–C1 English news article ChatGPT-4; Ferris & Hedgeock’s (2023) analytical rubric face-to-face; grammar/structure focus Feedback preference; Factors influencing preference Preference: Teacher + AI Collaborative > Native Teacher > ChatGPT (unaware) > Non-native Teacher > ChatGPT (aware). Abdi et al. (2025) USA Quasi-experimental; 2 weeks 166 ESL undergraduates; B2 level. Literacy narrative ChatGPT-4; Rubric prompt Individualized; emphasizing grammar, vocabulary, organization Accuracy;syntactic complexity; holistic quality Teacher: accuracy/lexicon; AI: syntax; Similar: overall quality. Yu et al. (2025) China Experimental design; 4 weeks 60 high school EFL students; CSE Level 4 Two argumentative essays ChatGPT; standardized prompt Direct and indirect comments; Notable inter-teacher variation Feedback traits; student uptake rates; revision AI: More comprehensive feedback; Teacher: Higher uptake rate, context-aware language. Topuz et al. (2025) Turkey Descriptive; NR 35 undergraduate EFL students; Advanced Three writing genres ChatGPT-3.5; Rubric prompts Dual trained raters; Analytical rubric Human vs AI essay scoring AI more consistent but stricter scorer; Significant human-AI score differences. Alnemrat et al. (2025) Jordan Quasi-experimental; over 1 week 120 undergrad EFL students; NR One argumentative essay ChatGPT-4.0; Rhetorical prompts Individualized; handwritten Writing performance, improvement gains; proficiency level Both AI and teacher feedback significantly improved writing scores. Zou et al. (2025) China Comparative study; around 1 month 20 Chinese undergraduates; CEFR B1–B2 Argumentative essay ChatGPT-4; Play the teacher's prompt Experienced teachers Uptake rate; revision success; perceptions; feedback focus Teacher: Higher uptake/success rate; AI: organizational; Students preferred teacher but valued AI support. Makwana (2025) India Mixed-methods; one semester First-semester postgraduate ESL students; NR 30 graded tasks Gemini Advanced; Trained on CEFR & IELTS rubrics prompt Personalized; contextualized Writing Performance; Student Perceptions; Feedback Quality AI & teachers: equally effective, complementary AI: grammar/vocab; Teacher: organization/coherence. Aljasser (2025) Saudi Arabia Mixed-methods; two semesters 112 Saudi EFL learners; C2 (CEFR) Four academic writing assignments ChatGPT-4; No specialized prompt design Rubric-based; personalized; detailed Perceived effectiveness, clarity, usefulness, personalization et al. Teacher: personalization, depth, motivation, and content/organization; AI: immediacy/accessibility and grammar/sentence structure. Escalante et al. (2023) USA Longitudinal quasi-exp; 6 weeks) 91 ENL students; ≥B1 (CEFR) Weekly 300 words academic paragraphs GPT-4.0; detailed task-specific prompt Face-to-face tutoring sessions; 30 mins weekly writing scores; student satisfaction, clarity, helpfulness, preference No outcome difference; AI: clarity; Teacher: interaction; Equal preference. Asadi et al. (2025) Iran Mixed-methods; 12 sessions 68 learners; B1-B2 (CEFR) IELTS Task 2 argumentative essays ChatGPT; teachers crafted prompts Traditional; personalized comments teacher-crafted prompts; overall score Teacher-crafted prompts; Synergy effective but requires oversight to mitigate overreliance risks. Notes : RCT = Randomized Controlled Trial, CAF = Complexity, Accuracy, Fluency, ENL = English as a New Language, CEFR = Common European Framework of Reference, IELTS = International English Language Testing System, GRE = Graduate Record Examinations, RM-ANOVA = Repeated Measures ANOVA Differential Efficacy of Feedback Modalities on Writing Outcomes Speaking of the writing development, a regular pattern of the difference could be observed among the feedback modalities (Table 2 ). In particular, AI-generated feedback proved to be more effective when it comes to fixing surface-level linguistic aspects that include grammatical accuracy, vocabulary diversity, and syntactic complexity, with several studies referring to the fact that AI was more effective at correcting discrete errors and expanding lexical resources (Ramazani et al., 2025; Aljasser, 2025; Cao et al., 2024). Contrastingly, the teacher feedback was found better to develop higher order writing skills, such as argument, structural organization, and coherence on a global scale. The distinctive skill of the teacher in encouraging the critical thinking and logical progression, as observed in research, was always a particular feature that cannot be assessed by AI contextually (Solovey, 2024; Zhang et al., 2025; Soori et al., 2025). Additionally, although both AI and teacher feedback might provide the same improvement in overall quality of writing (Abdi et al., 2025), teacher feedback tended to produce a more successful levels of student uptake and more extensive revisions, suggesting a higher level of cognitive processing of a human-expert input (Yu et al., 2025; Zou et al., 2025). Most importantly, as one of the strongest results throughout the systematic review, hybrid feedback models (i.e. combining AI and teacher feedback), by compensating the optimization of AI with the salary of the teacher time and the expertise) were more effective than either of the two single-source models in all cases by guaranteeing a strategic use of AI in error correction and retaining the time and the expertise of the teacher to provide complex, meaning-level feedback (Soori et al., 2025; Zhang et al., 2025; Asadi et al., 2025). Learner Perceptions, Preferences, and Affective Responses to AI and Teacher Feedback The interactions between practicability, credibility and affective experiences as perceptions of learners showed a complex interplay between these factors. The preferences of feedback did not remain the same and depended on the situation and the awareness of the learners about the sources of feedback. There was a strong hierarchy of preference in the study by Choi (2024): hybrid teacher-AI feedback received the highest ratings and native-speaker feedback and using ChatGPT without understanding it were slightly less preferred than non-native teacher feedback. According to this trend, learners still appreciate the authority and the fuzziness of human expertise and are also sufficiently aware of the convenience and effectiveness of AI whenever its participation is smoothly incorporated into the learning process. Perception of the two feedback modalities had further differences dependent on psychological and affective responses. The non-judgmental and never-available aspect of AI-generated feedback led to decreased writing anxiety (Wang, 2024) since it offered a low-stakes environment where more writing is practiced. Contrarily, teacher feedback was perceived as less personalized, less automatic, and more authoritative, as it fostered better confidence in academic writing in the long term and high relational support (Aljasser, 2025). Mode viewed strengths also conformed to its features of functionality. Students always appreciated AI in terms of urgency, 24 − 7 accessibility, and overall focus on grammar and lexical problems. Teacher feedback, in turn, was valued due to its situational understanding, its ability to make guesses at the intent of the writer, and the ability to provide encouraging, prospective advice beyond the superficial correction (Escalante et al., 2023). These findings combined show that the AI- and teacher-based feedback provided by the learner side can be seen as complements and that psychological factors should be taken into account in the feedback design. Discussion The findings of this systematic review clarify the complex interaction between generative AI and teacher feedback in EFL writing pedagogy, and not limited to the dichotomous comparisons alongside highlighting the synergies of both. This analysis, which synthesizes evidence on 17 studies, is consistent with recent empirical studies, including the study concerning the use of ChatGPT as a feedback provider in a variety of EFL settings (Choi, 2024). The identified bifurcation, the mastery of AI tools to ensure surface-level features such as grammar, vocabulary, and syntax makes the tools conform to the cognitive load theory according to which the accuracy of algorithmic systems prevents the unnecessary creation of cognitive load and allows students to focus on the content refinement (Sweller, 1988). As an example, an empirical study has found that AI applications, such as ChatGPT, can benefit EFL learners in their writing by merely providing specific corrections, but not by making revisions that go deeper (Ramazani et al., 2025). Nevertheless, there is also counter-evidence, including the AI being better than teachers in the reduction of writing anxiety (Wang, 2024), making affective areas the possible assets of AI. Teacher feedback, in contrast, proves to be more effective at developing higher-order skills, such as argumentation, organization, and coherence, through the use of contextual understanding and promotion of processes of metacognition. This is in line with sociocultural theory (Vygotsky, 1978) because teachers support meaning-with dialogue guidance that enhances meaningful uptake as observed in comparative research studies where teacher assistance was better at argumentative writing quality when compared to AI (Alnemrat et al., 2025). Nonetheless, the most crucial uncovering of the review is the dominance of hybrid models, which combine the speed of AI to provide mechanical fixes with the proficiencies of the teachers to induce a holistic development that provides higher outcomes of writing proficiency and revision depth (e.g., hybrid groups had higher gains in overall scores in various studies such as Soori et al., 2025). The scoping reviews that support this hybrid superiority presented the collaboration of AI and teachers as the most effective representation of L2 feedback, especially in the EFL setting (Zhang et al., 2025). However, this hope should be considered in moderation: hybrids encounter such issues as barriers in teacher training and the presence of biases in the artificial intelligence that work with non-Western languages, which may further contribute to inequities in the diverse EFL context (Yu et al., 2025). The preferences of learners as expressed in their perceptions are dynamic and situational with a tendency of leaning towards hybrids due to their combination of immediacy and authenticity. The practice based on low stakes which alleviates anxiety with the help of AI proves beneficial to the theory of self-determination (Deci and Ryan, 2020), whereas positive feedback expressed by teachers develops confidence and motivation over the long term (Wang, 2024). The trust in human aspects is additionally influenced by culture peculiarities that are common in various EFLs and are identified in systematic studies of GenAI in language classes (Zou et al., 2025). All of these discoveries add to the pre-AI feedback scholarship (Hyland and Hyland, 2006) and they place importance on the ability of GenAI to democratise access in resource scarce settings that also calls to the standardisation of approaches in SLA studies (Chapelle, 2001). Limitations This geographical sampling of East Asia, Middle East and Americas, underrepresenting all of Africa and Europe may reduce the generalizability of the findings and skew the insights on knowledge of cultures with stronger power distance, in which the teacher power is more promoted (Hofstede, 2001). Long-term effect assessment is confounded by the variability in the duration of intervention (weeks to semesters), which promotes short-term biases. Self-reported data is prone to social desirability, and methodological heterogeneity did not allow meta-analysis and instead required narrative synthesis. The non-native English situation with the AI technology may widen the digital gaps and discrimination issues when there is an ethical inefficiency or overtrust. The fast development of AI causes the risk of obsolescence (Escalante et al., 2023). These restrictions should be approached with reservations, and more searches and various sampling should be used in future studies. Implications and Future Directions Pedagogically, These possibilities pedagogically propose the framework of hybrids in teaching EFL, which offers AI, preliminary error correction to reduce the workload of teachers in their tasks and preserving human resources to provide complex scaffolding (Asadi et al., 2025). This method can potentially scale in crowded classrooms, although it needs to take care of such issues as educator education in timely engineering as well as the reduction of AI biases so that it is a fair addition. Policymakers are supposed to put a priority on AI literacy programs in order to achieve equal access and avoid addiction. Future studies should use longitudinal designs to assess lasting effects, diversify sample by demographic and region and test AI tools of the next generation against developing teacher strategies. More sophisticated models, like the structural equation modeling of psychological constructs like motivation may enhance the knowledge regarding preference mechanisms (Aljasser, 2025). It is also crucial to carry out cross-cultural validations and ethical questions of the role of GenAI in the development of critical thinking (Abdi et al., 2025). Conclusion This systematic review, synthesizing insights from 17 studies spanning 2023–2025, illuminates a strategic division of labor in EFL writing feedback: generative AI is more suitable in reducing anxiety and surface-level correction, and teacher intervention is more suitable in higher-order writing skills, such as argumentation and coherence. Combining the effectiveness of AI with the human expertise in the form of the hybrid models appears to be the best solution, as it improves the quality of writing, satisfaction of the learners, and mental performance. Practically, the educators are expected to put in place the tiered methodology of AI in performing first mechanical corrections and teachers in performing contextual corrections with integrative engineering and bias audit mechanisms in place to make equity a reality in various EFL situations. This instructional system facilitates learning and avoids excessive dependence on technology. Supported by the recent evidence, the given review recommends consideration of technology-based pedagogies to empower worldwide EFL learners in dynamic learning settings. It encourages teachers and decision makers to consider the ethical use of AI by integrating AI into the education system to close gaps in teaching writing across the globe. Declarations Author Contribution Author Contributions All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by Y. Zhao and D. Li, who independently screened 2,623 records, extracted data based on the PICOS framework, and assessed methodological quality using MMAT 2018. G. The first draft of the manuscript was written by Y. Zhao. All authors commented on previous versions of the manuscript, refined the thematic synthesis, and read and approved the final manuscript. Data Availability All data and materials supporting the findings of this study are available in the Supplementary Materials accompanying this manuscript. These include the PRISMA 2020 checklist and flow diagram, full database search strategies, screening logs with inclusion and exclusion criteria, extracted records, data extraction forms, and synthesis tables. The materials are provided to ensure transparency and reproducibility of the systematic review process. This article does not contain any studies with human participants performed by any of the authors References Abdi Tabari M, Kushki A, Wang Y (2025) Comparing the effects of teacher- and AI-mediated corrective feedback on accuracy, complexity, and quality in L2 written narratives. Comput Assist Lang Learn 1–24. https://doi.org/10.1080/09588221.2025.2561608 Aljasser A (2025) Investigating EFL students' perceptions of feedback: a comparative study of instructor and ChatGPT-generated responses in academic writing. Education and Information Technologies . Advance online publication. https://doi.org/10.1007/s10639-025-13649-8 Alnemrat A, Aldamen H, Almashour M, Al-Deaibes M, AlSharefeen R (2025) AI vs. teacher feedback on EFL argumentative writing: a quantitative study. Frontiers in Education , 10 . https://doi.org/10.3389/feduc.2025.1614673 Asadi M, Ebadi S, Mohammadi L (2025) The impact of integrating ChatGPT with teachers’ feedback on EFL writing skills. Think Skills Creativity 56:101766. https://doi.org/10.1016/j.tsc.2025.101766 Barrot JS (2023) Using ChatGPT for second language writing: Pitfalls and potentials. Assess Writ 57:100745. https://doi.org/10.1016/j.asw.2023.100745 Bitchener J, Ferris DR (2012) Written Corrective Feedback in Second Language Acquisition and Writing. Routledge. https://doi.org/10.4324/9780203832400 Cao S, Zhong L (2023) Exploring the effectiveness of ChatGPT-based feedback compared with teacher feedback and self-feedback: Evidence from Chinese to English translation [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2309.01645 Chapelle CA (2001) Computer Applications in Second Language Acquisition. Cambridge University Press, Cambridge Choi W (2025) Exploring high school students' preferences for English writing feedback: Non-native English teacher vs. native English teacher vs. ChatGPT. SAGE Open 15(4):1–16. https://doi.org/10.1177/21582440251383693 Cobb T (2008) Commentary: Response to mcquillan and krashen (2008). Language Learning & Technology , 12 (1), 109–114. https://doi.org/10.64152/10125/44134 Crowe M, Sheppard L (2011) A review of critical appraisal tools show they lack rigor: Alternative tool structure is proposed. J Clin Epidemiol 64(1):79–89 Derakhshan A (2025) EFL students’ perceptions about the role of generative artificial intelligence (GAI)-mediated instruction in their emotional engagement and goal orientation: A motivational climate theory (MCT) perspective in focus. Learn Motiv 90:102114. https://doi.org/10.1016/j.lmot.2025.102114 Escalante J, Pack A, Barrett A (2023) AI-generated feedback on writing: Insights into efficacy and ENL student preference. Int J Educational Technol High Educ 20(1):57. https://doi.org/10.1186/s41239-023-00425-2 Ferris D, Hyland K, Hyland F (2006) Does error feedback help student writers? New evidence on the short- and long-term effects of written error correction. Feedback in Second Language Writing: Contexts and Issues. Cambridge University Press, Cambridge, pp 81–104 Henderson M, Bearman M, Chung J, Fawns T, Shum SB, Matthews KE, De Mello Heredia J (2025) Comparing Generative AI and teacher feedback: student perceptions of usefulness and trustworthiness. Assess Evaluation High Educ 1–16. https://doi.org/10.1080/02602938.2025.2502582 Hofstede G (2001) Culture's consequences: Comparing values, behaviors, institutions and organizations across nations. Sage Hyland K, Hyland F (2006) Feedback on second language students’ writing. Lang Teach 39(2):83–101. 10.1017/S0261444806003399 Lo N, Chan S, Wong A (2025) Evaluating teacher, AI, and hybrid feedback in English Language Learning: Impact on student motivation, quality, and performance in Hong Kong. SAGE Open 15(3). https://doi.org/10.1177/21582440251352907 Makwana V (2025) A Comparative Analysis of AI-Powered and Teacher-Led Feedback: Investigating Student Perceptions and Writing Performance. J Engl Lang Teach 67(1):3–12. https://journals.eltai.in/jelt/article/view/JELT670102 Mekheimer M (2025) Generative AI-assisted feedback and EFL writing: A study on proficiency, revision frequency and writing quality. Discover Educ 4(1). Article 170. https://doi.org/10.1007/s44217-025-00602-7 Muñoz Muñoz BC, Nassaji H, Bello Carrillo FI (2025) ChatGPT-generated versus human direct corrective feedback on L2 writing. System 134:103805. https://doi.org/10.1016/j.system.2025.103805 Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, Shamseer L, Tetzlaff JM, Akl EA, Brennan SE, Chou R, Glanville J, Grimshaw JM, Hróbjartsson A, Lalu MM, Li T, Loder EW, Mayo-Wilson E, McDonald S, Moher D (2021) The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 372:n71. https://doi.org/10.1136/bmj.n71 Pluye P, Hong QN (2014) Combining the power of stories and the power of numbers: Mixed Methods Research and Mixed Studies Reviews. Annu Rev Public Health 35:29–45 Ramazani A, Bijani H, Oroji MR (2025) Comparative analysis of AI vs. human feedback effects on IELTS candidates' writing performance. J Foreign Lang Teach Translation Stud 10(1):17–40. https://doi.org/10.22034/efl.2025.493559.1334 Ryan RM, Deci EL (2020) Intrinsic and extrinsic motivation from a self-determination theory perspective: Definitions, theory, practices, and future directions. Contemp Educ Psychol 61:101860. https://doi.org/10.1016/j.cedpsych.2020.101860 Shen L, Wang S, Xin Y (2025) EFL students’ writing engagement and AI attitude in GenAI-assisted contexts: A mixed-methods study grounded in SDT and TAM. Learn Motiv 92:102168. https://doi.org/10.1016/j.lmot.2025.102168 Soloveyo OZ (2024) Comparing peer, ChatGPT, and teacher corrective feedback in EFL writing: Students' perceptions and preferences. Technol Lang Teach Learn 6(3):1–23. https://doi.org/10.29140/tltl.v6n3.1482 Soori A, Khojasteh L, Javed F (2025) Comparing teacher e-feedback, AI feedback, and hybrid feedback in enhancing EFL writing skills. Technol Lang Teach Learn 7(3). Article 102626. https://doi.org/10.29140/tltl.v7n3.102626 Sweller J (1988) Cognitive load during problem solving: Effects on learning. Cogn Sci 12(2):257–285. https://doi.org/10.1207/s15516709cog1202_4 Topuz AC, Yıldız M, Taşlıbeyaz E, Polat H, Kurşun E (2025) Is generative AI ready to replace human raters in scoring EFL writing? Comparison of human and automated essay evaluation. Educational Technol Soc 28(3):36–50. https://doi.org/10.30191/ETS.202507_28(3).SP04 Vygotsky LS (1978) Mind in society: The development of higher psychological processes. Harvard University Press. https://www.hup.harvard.edu/catalog.php?isbn=9780674576292 Wang D (2024) Teacher- versus AI-generated (Poe application) corrective feedback and language learners' writing anxiety, complexity, fluency, and accuracy. Int Rev Res Open Distrib Learn 25(3):37–56. https://doi.org/10.19173/irrodl.v25i3.7646 Warschauer M, Tseng W, Yim S, Webster T, Jacob S, Du Q, Tate T (2023) The affordances and contradictions of AI-generated text for writers of english as a second or foreign language. J Second Lang Writ 62:101071. https://doi.org/10.1016/j.jslw.2023.101071 Yang Y, Yao S (2025) Beyond linguistic proficiency: emotional, cognitive, and cultural pathways of Generative AI in EFL Well-Being—A Technology‐Humanities‐Culture Synthesis. Int J Appl Linguistics. https://doi.org/10.1111/ijal.12832 Yu H, Xie Q (2025) Generative AI vs. teachers: Feedback quality, feedback uptake, and revision. Lang Teach Res Q 47:113–137. https://doi.org/10.32038/ltrq.2025.47.07 Zhang Z (2020) Engaging with automated writing evaluation (AWE) feedback on L2 writing: Student perceptions and revisions. Assess Writ 43:100439. https://doi.org/10.1016/j.asw.2019.100439 Zhang Z, Aubrey S, Huang X, Chiu TKF (2025) The role of generative AI and hybrid feedback in improving L2 writing skills: a comparative study. Innov Lang Learn Teach 1–19. https://doi.org/10.1080/17501229.2025.2503890 Zou S, Guo K, Wang J, Liu Y (2025) Investigating students’ uptake of teacher- and ChatGPT-generated feedback in EFL writing: a comparison study. Comput Assist Lang Learn 1–30. https://doi.org/10.1080/09588221.2024.2447279 Additional Declarations No competing interests reported. Supplementary Files SupplementaryMaterials.rar Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 17 May, 2026 Reviewers agreed at journal 02 May, 2026 Reviewers agreed at journal 01 May, 2026 Reviews received at journal 01 May, 2026 Reviewers agreed at journal 19 Apr, 2026 Reviews received at journal 08 Mar, 2026 Reviewers agreed at journal 29 Jan, 2026 Reviewers agreed at journal 22 Jan, 2026 Reviewers invited by journal 22 Jan, 2026 Editor assigned by journal 22 Jan, 2026 Editor invited by journal 08 Jan, 2026 Submission checks completed at journal 06 Jan, 2026 First submitted to journal 06 Jan, 2026 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-8386846","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Systematic Review","associatedPublications":[],"authors":[{"id":579688000,"identity":"70aa6d79-3be5-4a1d-8e0a-b80b6c4169dc","order_by":0,"name":"Yanmei Zhao¹","email":"","orcid":"","institution":"Yuxi Normal University","correspondingAuthor":false,"prefix":"","firstName":"Yanmei","middleName":"","lastName":"Zhao¹","suffix":""},{"id":579688001,"identity":"9248dd0f-ff4d-49a8-a7e3-f74ea754587b","order_by":1,"name":"Dandan Li²","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA80lEQVRIiWNgGAWjYDACCQjFz8/efAAqlECcFsmZPcdgSonVsuFGjgFxWvhnNx/7XNl2WMLgRs43ad4dhxn42YF6f+7AY8mdY8kzz7alSUieebtNcuaZwwySPW8MGHvP4NZiIJFjzNjYZlPHdzx3m8THtsMMQOsMmBnb8GnJ/wzUIiHBcCDnmUQiUIs9YS05zCBbJARO5LBBbJEgoEXiRpoxY8M5oF96jhlbzmxL55E486zgYC8eLfwzkh8zNpQdlgBG5cPbvG3WcvztyRsf/MSjBQwY2cAUCyiOeECsAwQ0AMEfMMn8gbDKUTAKRsEoGIkAAEKgUPKmqKH4AAAAAElFTkSuQmCC","orcid":"","institution":"Wuhu Vocational Technical University","correspondingAuthor":true,"prefix":"","firstName":"Dandan","middleName":"","lastName":"Li²","suffix":""}],"badges":[],"createdAt":"2025-12-17 14:38:24","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8386846/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8386846/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":101253519,"identity":"da587e60-544a-4d6d-9725-06abaef25aa4","added_by":"auto","created_at":"2026-01-27 18:14:26","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":179997,"visible":true,"origin":"","legend":"\u003cp\u003ePRISMA Flow Chart of the Studies Included in the Systematic Review\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8386846/v1/8f8a408dc435df9ce6064855.png"},{"id":101297702,"identity":"5cee4efa-17dd-4659-b923-6e51f594ad2b","added_by":"auto","created_at":"2026-01-28 09:28:40","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1095298,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8386846/v1/b41a3a39-3e13-4646-bae1-5d124f25e71e.pdf"},{"id":101253521,"identity":"277d1d7b-e1f9-4bab-9fed-c03f6ae9301e","added_by":"auto","created_at":"2026-01-27 18:14:27","extension":"rar","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":18372159,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterials.rar","url":"https://assets-eu.researchsquare.com/files/rs-8386846/v1/732bf3ef3439b1dad411759b.rar"}],"financialInterests":"No competing interests reported.","formattedTitle":"A Systematic Comparative Review of Generative AI vs. Teacher Feedback in EFL Writing: Learner Perceptions, Preferences, and Psychological Factors","fulltext":[{"header":"Introduction","content":"\u003cp\u003eWritten corrective feedback (WCF) has been one of the foundations of second language (L2) writing education, which is a critical factor in achieving linguistic accuracy and proficiency writing (Bitchener and Ferris, 2012; Hyland and Hyland, 2006). This is especially important in English as a Foreign Language (EFL) settings, when the availability of high-quality feedback is also usually very restricted. Although conventional teacher feedback is being cherished due to the ability to identify the higher-order issues and to be sensitive to context-based concerns (Ferris, 2006; Hyland, 2013), it is often undermined by the practical factors, including a high number of students in a classroom and the workload of a teacher (Zhang, 2020).\u003c/p\u003e \u003cp\u003eThis landscape can be changed with the emergence of generative artificial intelligence (AI), namely large language models such as ChatGPT. These systems offer instant, scalable, and linguistically advanced feedback, which may result in human-provided WCF overcoming some of its essential constraints (Barrot, 2023; Warschauer et al., 2023; Mekheimer, 2025). Participating in contrast to the previous automated writing assessors that tried to detect the errors only on the surface (Cobb, 2008), the modern generative AI can provide detailed feedback on the content, structure, and the rhetoric, and in some cases, it can even excel at human tasks. These potentials are highlighted by research on emerging comparative studies, which determine that GenAI is effective in alleviating writing anxiety and boosting lexical resources (Wang, 2024; Ramazani et al., 2025). Nevertheless, one of the constant conclusions of the recent literature is that it is relatively weak in building strong cognitive involvement, effective acquisition, and metacognitive awareness relative to the level of fine guidance of instructors (Mu\u0026ntilde;oz Mu\u0026ntilde;oz et al., 2025; Yu and Xie, 2025).\u003c/p\u003e \u003cp\u003eNevertheless, even with this quickly growing body of empirical research since 2023 the subject matter remains counterintuitive, conclusively, inconsistent. There is also no meta-analysis to date (Wang and Dang, 2024), the findings on qualitative level are mostly cultural siloed (Yang and Yao, 2025; ), and the most essential aspects such as perceptions, preferences, and psychological reactions of learners (e.g., a sense of autonomy, relational trust, writing anxiety) that ultimately mediate feedback uptake, have never been synthesised (Henderson et al., 2025; Shen et al., 2025). Moreover, even though hybrid human-AI feedback models enjoy a growing body of support (Zhang et al., 2025), they have minimal empirical basis and lack contextual differentiation (Derakhshan, 2025; Lo et al., 2025). There is therefore an urgent need to have a detailed systematic synthesis to justify these opposing results, consider contextual variables, and evaluate the empirical rationale upon the integration of the two information sources of feedback.\u003c/p\u003e \u003cp\u003eThis systematic review intends to bring together the still-expanding fragmented data on comparative research in the field of generative AI and teacher feedback in EFL writing guided by the self-determination theory (Ryan and Deci, 2020) and the cognitive load theory. It aims at fulfilling the following purposes: firstly, to critically scrutinize and resolve, the discrepant empirical evidence on their varied effects on the development of L2 writing, especially on complexity, accuracy, fluency, and overall quality; second, to unite qualitative and quantitative results concerning how learners perceive, how they like, and how they are psychologically inclined to respond to feedback, and third, to weigh up the empirical evidence upon emergent hybrid feedback models and to suggest a coherent and culturally aware framework through which they should be taught.\u003c/p\u003e"},{"header":"Methodology","content":"\u003cp\u003eInformation Sources and Search Strategy\u003c/p\u003e \u003cp\u003eA systematic literature search was performed across several major electronic databases, including Google Scholar, Web of Science, ScienceDirect, SpringerLink, and SAGE Journals. To capture the entire body of relevant literature since the emergence of generative AI, the search covered all publications from the inception of each database up to the date the search was conducted (November 2025).\u003c/p\u003e \u003cp\u003eThe search strategy was designed to be broad and inclusive, applying the following Boolean query to the \u0026ldquo;All Fields\u0026rdquo; metadata where possible: ( (EFL OR \"English as a Foreign Language\" OR \"English learner\" OR \"L2 writer\" OR \"second language writ\" OR \"non-native speaker\") ) AND ( (\u0026ldquo;Generative AI\u0026rdquo; OR \u0026ldquo;Generative Artificial Intelligence\u0026rdquo; OR \u0026ldquo;ChatGPT\u0026rdquo; OR \u0026ldquo;GPT-3\u0026rdquo; OR \u0026ldquo;GPT-4\u0026rdquo; OR \u0026ldquo;large language model\u0026rdquo; OR \u0026ldquo;LLM\u0026rdquo; OR \u0026ldquo;AI-generated feedback\u0026rdquo; OR \u0026ldquo;AI writing tool\u0026rdquo;) ) AND ( (\u0026ldquo;teacher feedback\u0026rdquo; OR \u0026ldquo;instructor feedback\u0026rdquo; OR \u0026ldquo;human feedback\u0026rdquo;) ).\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eEligibility Criteria\u003c/h2\u003e \u003cp\u003eStudies were selected based on the pre-defined PICOS framework: Population (P): Secondary or tertiary-level EFL learners. Intervention (I): Writing feedback generated by generative AI models (e.g., ChatGPT). Comparator (C): Feedback provided by teachers or human instructors. Outcomes (O): Comparative measures of writing performance (e.g., CAF, rubric scores), feedback uptake, or learner perceptions. Study Design (S): Empirical comparative studies (e.g., RCTs, quasi-experimental, mixed-methods).\u003c/p\u003e \u003cp\u003eExclusion criteria were: Non-empirical publications (e.g., reviews, theoretical papers). Studies without a direct comparison between generative AI and teacher feedback. Articles using non-generative AI systems (e.g., traditional AWE). Non-English publications.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eStudy Selection Process\u003c/h3\u003e\n\u003cp\u003eThe selection of the study adhered to PRISMA flow diagram (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). A total of 2,623 records were initial; 500 records were duplicates. Titles and abstracts of 2123 unique records were screened with 1910 being excluded. The rest of the 213 articles were evaluated and 196 articles were eliminated because of certain reasons: lack of direct comparison (n\u0026thinsp;=\u0026thinsp;121), the employment of non-generative AI (n\u0026thinsp;=\u0026thinsp;48), and lack of data (n\u0026thinsp;=\u0026thinsp;26). An ultimate sample size of 17 studies was added. This was done by two reviewers working individually and any difference was overcome by consensus.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eData Extraction\u003c/h3\u003e\n\u003cp\u003eA systematic data extraction process was done in order to systematically identify the most important characteristics of the included studies, pointing towards a pre-defined checklist in accordance with the PRISMA guidelines (Page et al., 2021). Two reviewers (Y. Zhao and D. Li) independently identified data on four core domains (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) basic study information (authors, year, country, design, duration); (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) participant profile (sample size, level of proficiency); (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) intervention (AI tool and prompt design, teacher feedback characteristics, writing task); (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e) key findings of the AI-teacher comparison. Minor differences among the reviewers were addressed by consensus discussion, and when the facing out of an issue was necessary, a third senior researcher was consulted.\u003c/p\u003e\n\u003ch3\u003eQuality Assessment\u003c/h3\u003e\n\u003cp\u003eAll the studies (n\u0026thinsp;=\u0026thinsp;18) included were evaluated with the Mixed Methods Appraisal Tool (MMAT) 2018 (Crowe et al., 2011; Pluye and Hong, 2014), which is a well-validated evaluation instrument of the quality of study designs (mixed methods, quantitative descriptive, non-randomized, RCT, and qualitative studies). The MMAT test included two obligatory screening questions (S1-S2) and criteria that depended on the research design (1.1\u0026ndash;5.5). All three researchers had to train on MMAT 2018 before the formal assessment to ascertain their similarity in the definition of criteria and standards on how to pass marks. The two scholars (Y. Zhao and D. Li) independently rated each study. Poor ratings in terms of quality were sorted out through group discussion.\u003c/p\u003e \u003cp\u003eThe results of quality assessment showed that the quality of methodological approaches of all the considered studies was satisfactory. In particular, 15 articles (83.3% of all) were considered to be of high quality, these articles met all possible MMAT criteria. The rest of the studies (16.7% of all) were classified as having medium quality; they did not fail to satisfy the two screening questions or the majority of the design-specific criteria, but they were somehow limited (e.g., not a randomized study, not well integrated quantitative and qualitative findings in mixed methods studies). None of the studies were considered to be of low quality, which ensured that the used research materials were credible to carry out further analysis (See Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\n\u003ch3\u003eData Synthesis\u003c/h3\u003e\n\u003cp\u003eSince the extent of heterogeneity of experimental designs and outcomes measures of the studies included was very high, a narrative synthesis methodology was used in analyzing the data. Quantitative findings were summarized and shown descriptively and qualitative findings analyzed using a thematic analysis. This was sought to determine convergent and divergent trends of the comparative effectiveness of AI and teacher feedback, learners perceptions and preferences.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSummary of methodological quality assessment\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"29\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c13\" colnum=\"13\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c14\" colnum=\"14\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c15\" colnum=\"15\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c16\" colnum=\"16\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c17\" colnum=\"17\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c18\" colnum=\"18\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c19\" colnum=\"19\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c20\" colnum=\"20\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c21\" colnum=\"21\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c22\" colnum=\"22\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c23\" colnum=\"23\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c24\" colnum=\"24\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c25\" colnum=\"25\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c26\" colnum=\"26\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c27\" colnum=\"27\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c28\" colnum=\"28\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c29\" colnum=\"29\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eFirst Author, Year\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eSCREENING QUESTIONS\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"5\" nameend=\"c8\" namest=\"c4\"\u003e \u003cp\u003e1. QUALITATIVE STUDIES\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"5\" nameend=\"c13\" namest=\"c9\"\u003e \u003cp\u003e2. RANDOMIZED CONTROLLED TRIALS\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"5\" nameend=\"c18\" namest=\"c14\"\u003e \u003cp\u003e3. NON-RANDOMIZED STUDIES\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"5\" nameend=\"c23\" namest=\"c19\"\u003e \u003cp\u003e4. QUANTITATIVE DESCRIPTIVE STUDIES\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"5\" nameend=\"c28\" namest=\"c24\"\u003e \u003cp\u003e5. MIXED METHODS STUDIES\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c29\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eQuality Rating\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eS1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eS2\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.2\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.3\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.4\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1.5\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003e2.1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003e2.2\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003e2.3\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e \u003cp\u003e2.4\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c13\"\u003e \u003cp\u003e2.5\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c14\"\u003e \u003cp\u003e3.1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c15\"\u003e \u003cp\u003e3.2\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c16\"\u003e \u003cp\u003e3.3\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c17\"\u003e \u003cp\u003e3.4\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c18\"\u003e \u003cp\u003e3.5\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c19\"\u003e \u003cp\u003e4.1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c20\"\u003e \u003cp\u003e4.2\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c21\"\u003e \u003cp\u003e4.3\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c22\"\u003e \u003cp\u003e4.4\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c23\"\u003e \u003cp\u003e4.5\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c24\"\u003e \u003cp\u003e5.1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c25\"\u003e \u003cp\u003e5.2\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c26\"\u003e \u003cp\u003e5.3\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c27\"\u003e \u003cp\u003e5.4\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c28\"\u003e \u003cp\u003e5.5\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRamazani et al., 2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003ehigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSolovey, 2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003ehigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCao et al., 2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003ehigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSoori et al., 2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e\u0026times;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003emedium\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWang, 2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003ehigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eZhang et al., 2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u0026times;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e\u0026times;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003emedium\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMu\u0026ntilde;oz et al.,2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003ehigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eChoi, 2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003ehigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAbdi et al., 2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003ehigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYu et al., 2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003ehigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTopuz et al., 2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003ehigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAlnemrat et al., 2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003ehigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eZou et al., 2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003ehigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMakwana, 2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003ehigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAljasser, 2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003ehigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEscalante et al., 2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003ehigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAsadi et al., 2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c17\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c18\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c19\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c20\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c21\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c22\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c23\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c24\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c25\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c26\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c27\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c28\"\u003e \u003cp\u003e✔\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c29\"\u003e \u003cp\u003ehigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"29\"\u003e\u003cem\u003eN\u003c/em\u003eotes: S1\u0026thinsp;=\u0026thinsp;Are there clear research questions?; S2\u0026thinsp;=\u0026thinsp;Do the collected data allow to address the research questions?; 1.1\u0026thinsp;=\u0026thinsp;Is the qualitative approach appropriate?; 1.2\u0026thinsp;=\u0026thinsp;Are qualitative data collection methods adequate?; 1.3\u0026thinsp;=\u0026thinsp;Are findings derived from data?; 1.4\u0026thinsp;=\u0026thinsp;Is interpretation substantiated by data?; 1.5\u0026thinsp;=\u0026thinsp;Is there coherence between components?; 2.1\u0026thinsp;=\u0026thinsp;Is randomization appropriately performed?; 2.2\u0026thinsp;=\u0026thinsp;Are groups comparable at baseline?; 2.3\u0026thinsp;=\u0026thinsp;Are there complete outcome data?; 2.4\u0026thinsp;=\u0026thinsp;Are outcome assessors blinded?; 2.5\u0026thinsp;=\u0026thinsp;Did participants adhere to intervention?; 3.1\u0026thinsp;=\u0026thinsp;Are participants representative?; 3.2\u0026thinsp;=\u0026thinsp;Are measurements appropriate?; 3.3\u0026thinsp;=\u0026thinsp;Are there complete outcome data?; 3.4\u0026thinsp;=\u0026thinsp;Are confounders accounted for?; 3.5\u0026thinsp;=\u0026thinsp;Was intervention administered as intended?; 4.1\u0026thinsp;=\u0026thinsp;Is sampling strategy relevant?; 4.2\u0026thinsp;=\u0026thinsp;Is sample representative?; 4.3\u0026thinsp;=\u0026thinsp;Are measurements appropriate?; 4.4\u0026thinsp;=\u0026thinsp;Is risk of nonresponse bias low?; 4.5\u0026thinsp;=\u0026thinsp;Is statistical analysis appropriate?; 5.1\u0026thinsp;=\u0026thinsp;Is there adequate rationale for mixed methods?; 5.2\u0026thinsp;=\u0026thinsp;Are components effectively integrated?; 5.3\u0026thinsp;=\u0026thinsp;Are integrated outputs adequately interpreted?; 5.4\u0026thinsp;=\u0026thinsp;Are divergences between results addressed?; 5.5\u0026thinsp;=\u0026thinsp;Do components adhere to quality criteria of each method?; ✔=Yes; \u0026times;=NO; \u0026ldquo;-\u0026rdquo; indicates that the criterion is not applicable to the study design.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eCharacteristics of the Included Studies\u003c/h2\u003e \u003cp\u003eThe scope of geographical distribution of the 17 analyzed studies is quite broad: the studies were carried out in East Asian (China, Hong Kong, South Korea) and Middle Eastern (Iran, Israel, Jordan, Saudi Arabia, Turkey) settings as well as in South Asian (India) and American (USA, Chile) settings. Such a variety highlights the international concern about the field of research (See Table \u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The studies had a variety of methods used, such as randomized controlled trials (RCTs), quasi-experimental, mixed-methods, and qualitative in the methodological aspect. The length of the interventions was not exactly the same, as short, intensive investigations of a week length were also interspersed with longitudinal studies lasting an entire semester of academia. The participants were students with A2 proficiency to postgraduate students (C2 proficiency, most of them at the undergraduate level, i.e. B1-C1). ChatGPT (with versions GPT-3.5, GPT-4, and GPT-4 Turbo) was the most commonly used AI tool, and prompt design became one of the key aspects that determine the quality of a piece of AI-generated feedback. Teacher feedback was also described as individualized, context-dependent and usually on the higher-order issues in writing.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCharacteristics of Studies Included\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"9\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAuthors (Year)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCountry\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDesign; Duration\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eParticipants; Proficiency\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eWriting Task\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eAI Tool; Prompt Design\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eTeacher Feedback Characteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eCore Comparison Dimensions\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eKey Findings\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRamazani et al. (2025)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIran\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMixed-methods; 6 weeks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e100 IELTS candidates; B2-level\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eIELTS Writing Tasks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eChatGPT 3.0; IELTS-aligned prompts\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003ePersonalized annotations\u0026thinsp;+\u0026thinsp;consultations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eIELTS 4 criteria\u0026thinsp;+\u0026thinsp;learner perceptions\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eTeacher better for higher-order; AI for surface errors; Hybrid recommended\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSolovey (2024)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIsrael\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eQualitative; NR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e30 EFL students; B2 level\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eArgumentative essay\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eChatGPT; Checklist prompts\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eWritten corrective feedback\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003ePerceived effectiveness; feedback preferences\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eTeacher preferred for expertise/reliability; AI valued for speed/comprehensiveness.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCao et al. (2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eExperimenta; 2-week intervals\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e45 MTI students; TEM-8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eChinese-English translation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eChatGPT-4; standardized prompt\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eContextual feedback; translation logic; syntactic accuracy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eTranslation quality; linguistic features (lexicon/syntax/cohesion)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAI excels in lexical capacity; Teacher superior for syntax and overall quality; Blended approach recommended.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSoori et al. (2025)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIran\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRandomized controlled tria; 2.5 months\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e88 IELTS EFL learners; Intermediate\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e200 word IELTS Task 2 essays\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eChatGPT-4, Grammarly; prompts: IELTS-aligned\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eE-feedback; personalize; context-rich; higher-order skills\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eIELTS criteria; writing performance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eHybrid\u0026thinsp;\u0026gt;\u0026thinsp;AI or Teacher; AI\u0026thinsp;\u0026gt;\u0026thinsp;Teacher (vocab); Teacher\u0026thinsp;\u0026gt;\u0026thinsp;AI (higher-order); Students prefer hybrid.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWang (2024)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eQuasi-experimental; 14 weeks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e75 undergraduate EFL learners; NR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eWeekly course assignments\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePoe app; prompts: grammar/coherence\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eTraditional corrective\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eAnxiety; CAF\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAI\u0026thinsp;\u0026gt;\u0026thinsp;Teacher (anxiety, accuracy, fluency, complexity).\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eZhang et al. (2025)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eExperimental; 12 weeks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e60 Chinese EFL students; TOEFL\u0026thinsp;\u0026gt;\u0026thinsp;90, IELTS\u0026thinsp;\u0026gt;\u0026thinsp;7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eGRE Issue Writing Task\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGPT-4 Turbo; GRE rubrics\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003ePersonalized; supplements AI feedback; higher-order skills\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eGRE writing criteria; motivation; feedback perception\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eHybrid feedback superior to AI-only; AI for basic skills but limited in higher-order thinking.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMu\u0026ntilde;oz et al. (2025)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eChile\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eQuasi-experimental; Over 2 months\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e44 EFL students; CEFR A2\u0026ndash;B1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eIELTS-style opinion essays\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGPT-4o; Role-playing as English teacher\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDirect written corrective; Band scores\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eIELTS criteria\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eChatGPT\u0026thinsp;\u0026gt;\u0026thinsp;Teacher (all criteria); Large effect size for AI.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eChoi (2024)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSouth Korea\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMixed-methods; NR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e20 high school students; CEFR B2\u0026ndash;C1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eEnglish news article\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eChatGPT-4; Ferris \u0026amp; Hedgeock\u0026rsquo;s (2023) analytical rubric\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eface-to-face; grammar/structure focus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eFeedback preference; Factors influencing preference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003ePreference: Teacher\u0026thinsp;+\u0026thinsp;AI Collaborative\u0026thinsp;\u0026gt;\u0026thinsp;Native Teacher\u0026thinsp;\u0026gt;\u0026thinsp;ChatGPT (unaware)\u0026thinsp;\u0026gt;\u0026thinsp;Non-native Teacher\u0026thinsp;\u0026gt;\u0026thinsp;ChatGPT (aware).\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAbdi et al. (2025)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUSA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eQuasi-experimental; 2 weeks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e166 ESL undergraduates; B2 level.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eLiteracy narrative\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eChatGPT-4; Rubric prompt\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eIndividualized; emphasizing grammar, vocabulary, organization\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eAccuracy;syntactic complexity; holistic quality\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eTeacher: accuracy/lexicon; AI: syntax; Similar: overall quality.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYu et al. (2025)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eExperimental design; 4 weeks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e60 high school EFL students; CSE Level 4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eTwo argumentative essays\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eChatGPT; standardized prompt\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDirect and indirect comments; Notable inter-teacher variation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eFeedback traits; student uptake rates; revision\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAI: More comprehensive feedback; Teacher: Higher uptake rate, context-aware language.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTopuz et al. (2025)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTurkey\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDescriptive; NR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e35 undergraduate EFL students; Advanced\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eThree writing genres\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eChatGPT-3.5; Rubric prompts\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDual trained raters; Analytical rubric\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eHuman vs AI essay scoring\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAI more consistent but stricter scorer; Significant human-AI score differences.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAlnemrat et al. (2025)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eJordan\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eQuasi-experimental; over 1 week\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e120 undergrad EFL students; NR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eOne argumentative essay\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eChatGPT-4.0; Rhetorical prompts\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eIndividualized; handwritten\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eWriting performance, improvement gains; proficiency level\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eBoth AI and teacher feedback significantly improved writing scores.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eZou et al. (2025)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eComparative study; around 1 month\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e20 Chinese undergraduates; CEFR B1\u0026ndash;B2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eArgumentative essay\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eChatGPT-4; Play the teacher's prompt\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eExperienced teachers\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eUptake rate; revision success; perceptions; feedback focus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eTeacher: Higher uptake/success rate; AI: organizational; Students preferred teacher but valued AI support.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMakwana (2025)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIndia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMixed-methods; one semester\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFirst-semester postgraduate ESL students; NR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e30 graded tasks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGemini Advanced; Trained on CEFR \u0026amp; IELTS rubrics prompt\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003ePersonalized; contextualized\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eWriting Performance;\u003c/p\u003e \u003cp\u003eStudent Perceptions; \u003c/p\u003e \u003cp\u003eFeedback Quality\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAI \u0026amp; teachers: equally effective, complementary AI: grammar/vocab; Teacher: organization/coherence.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAljasser (2025)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSaudi Arabia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMixed-methods; two semesters\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e112 Saudi EFL learners; C2 (CEFR)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eFour academic writing assignments\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eChatGPT-4; No specialized prompt design\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eRubric-based; personalized; detailed\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003ePerceived effectiveness, clarity, usefulness, personalization et al.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eTeacher: personalization, depth, motivation, and content/organization; AI: immediacy/accessibility and grammar/sentence structure.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEscalante et al. (2023)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUSA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLongitudinal quasi-exp; 6 weeks)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e91 ENL students; \u0026ge;B1 (CEFR)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eWeekly 300 words academic paragraphs\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGPT-4.0; detailed task-specific prompt\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eFace-to-face tutoring sessions; 30 mins weekly\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003ewriting scores; student satisfaction, clarity, helpfulness, preference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eNo outcome difference; AI: clarity; Teacher: interaction; Equal preference.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAsadi et al. (2025)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIran\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMixed-methods; 12 sessions\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e68 learners; B1-B2 (CEFR)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eIELTS Task 2 argumentative essays\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eChatGPT; teachers crafted prompts\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eTraditional; personalized comments\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eteacher-crafted prompts; overall score\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eTeacher-crafted prompts; Synergy effective but requires oversight to mitigate overreliance risks.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"9\"\u003e\u003cem\u003eNotes\u003c/em\u003e: RCT\u0026thinsp;=\u0026thinsp;Randomized Controlled Trial, CAF\u0026thinsp;=\u0026thinsp;Complexity, Accuracy, Fluency, ENL\u0026thinsp;=\u0026thinsp;English as a New Language, CEFR\u0026thinsp;=\u0026thinsp;Common European Framework of Reference, IELTS\u0026thinsp;=\u0026thinsp;International English Language Testing System, GRE\u0026thinsp;=\u0026thinsp;Graduate Record Examinations, RM-ANOVA\u0026thinsp;=\u0026thinsp;Repeated Measures ANOVA\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eDifferential Efficacy of Feedback Modalities on Writing Outcomes\u003c/p\u003e \u003cp\u003eSpeaking of the writing development, a regular pattern of the difference could be observed among the feedback modalities (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). In particular, AI-generated feedback proved to be more effective when it comes to fixing surface-level linguistic aspects that include grammatical accuracy, vocabulary diversity, and syntactic complexity, with several studies referring to the fact that AI was more effective at correcting discrete errors and expanding lexical resources (Ramazani et al., 2025; Aljasser, 2025; Cao et al., 2024).\u003c/p\u003e \u003cp\u003eContrastingly, the teacher feedback was found better to develop higher order writing skills, such as argument, structural organization, and coherence on a global scale. The distinctive skill of the teacher in encouraging the critical thinking and logical progression, as observed in research, was always a particular feature that cannot be assessed by AI contextually (Solovey, 2024; Zhang et al., 2025; Soori et al., 2025). Additionally, although both AI and teacher feedback might provide the same improvement in overall quality of writing (Abdi et al., 2025), teacher feedback tended to produce a more successful levels of student uptake and more extensive revisions, suggesting a higher level of cognitive processing of a human-expert input (Yu et al., 2025; Zou et al., 2025).\u003c/p\u003e \u003cp\u003eMost importantly, as one of the strongest results throughout the systematic review, hybrid feedback models (i.e. combining AI and teacher feedback), by compensating the optimization of AI with the salary of the teacher time and the expertise) were more effective than either of the two single-source models in all cases by guaranteeing a strategic use of AI in error correction and retaining the time and the expertise of the teacher to provide complex, meaning-level feedback (Soori et al., 2025; Zhang et al., 2025; Asadi et al., 2025).\u003c/p\u003e \u003cp\u003eLearner Perceptions, Preferences, and Affective Responses to AI and Teacher Feedback\u003c/p\u003e \u003cp\u003eThe interactions between practicability, credibility and affective experiences as perceptions of learners showed a complex interplay between these factors. The preferences of feedback did not remain the same and depended on the situation and the awareness of the learners about the sources of feedback. There was a strong hierarchy of preference in the study by Choi (2024): hybrid teacher-AI feedback received the highest ratings and native-speaker feedback and using ChatGPT without understanding it were slightly less preferred than non-native teacher feedback. According to this trend, learners still appreciate the authority and the fuzziness of human expertise and are also sufficiently aware of the convenience and effectiveness of AI whenever its participation is smoothly incorporated into the learning process.\u003c/p\u003e \u003cp\u003ePerception of the two feedback modalities had further differences dependent on psychological and affective responses. The non-judgmental and never-available aspect of AI-generated feedback led to decreased writing anxiety (Wang, 2024) since it offered a low-stakes environment where more writing is practiced. Contrarily, teacher feedback was perceived as less personalized, less automatic, and more authoritative, as it fostered better confidence in academic writing in the long term and high relational support (Aljasser, 2025).\u003c/p\u003e \u003cp\u003eMode viewed strengths also conformed to its features of functionality. Students always appreciated AI in terms of urgency, 24\u0026thinsp;\u0026minus;\u0026thinsp;7 accessibility, and overall focus on grammar and lexical problems. Teacher feedback, in turn, was valued due to its situational understanding, its ability to make guesses at the intent of the writer, and the ability to provide encouraging, prospective advice beyond the superficial correction (Escalante et al., 2023). These findings combined show that the AI- and teacher-based feedback provided by the learner side can be seen as complements and that psychological factors should be taken into account in the feedback design.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe findings of this systematic review clarify the complex interaction between generative AI and teacher feedback in EFL writing pedagogy, and not limited to the dichotomous comparisons alongside highlighting the synergies of both. This analysis, which synthesizes evidence on 17 studies, is consistent with recent empirical studies, including the study concerning the use of ChatGPT as a feedback provider in a variety of EFL settings (Choi, 2024). The identified bifurcation, the mastery of AI tools to ensure surface-level features such as grammar, vocabulary, and syntax makes the tools conform to the cognitive load theory according to which the accuracy of algorithmic systems prevents the unnecessary creation of cognitive load and allows students to focus on the content refinement (Sweller, 1988). As an example, an empirical study has found that AI applications, such as ChatGPT, can benefit EFL learners in their writing by merely providing specific corrections, but not by making revisions that go deeper (Ramazani et al., 2025). Nevertheless, there is also counter-evidence, including the AI being better than teachers in the reduction of writing anxiety (Wang, 2024), making affective areas the possible assets of AI.\u003c/p\u003e \u003cp\u003eTeacher feedback, in contrast, proves to be more effective at developing higher-order skills, such as argumentation, organization, and coherence, through the use of contextual understanding and promotion of processes of metacognition. This is in line with sociocultural theory (Vygotsky, 1978) because teachers support meaning-with dialogue guidance that enhances meaningful uptake as observed in comparative research studies where teacher assistance was better at argumentative writing quality when compared to AI (Alnemrat et al., 2025). Nonetheless, the most crucial uncovering of the review is the dominance of hybrid models, which combine the speed of AI to provide mechanical fixes with the proficiencies of the teachers to induce a holistic development that provides higher outcomes of writing proficiency and revision depth (e.g., hybrid groups had higher gains in overall scores in various studies such as Soori et al., 2025). The scoping reviews that support this hybrid superiority presented the collaboration of AI and teachers as the most effective representation of L2 feedback, especially in the EFL setting (Zhang et al., 2025). However, this hope should be considered in moderation: hybrids encounter such issues as barriers in teacher training and the presence of biases in the artificial intelligence that work with non-Western languages, which may further contribute to inequities in the diverse EFL context (Yu et al., 2025).\u003c/p\u003e \u003cp\u003eThe preferences of learners as expressed in their perceptions are dynamic and situational with a tendency of leaning towards hybrids due to their combination of immediacy and authenticity. The practice based on low stakes which alleviates anxiety with the help of AI proves beneficial to the theory of self-determination (Deci and Ryan, 2020), whereas positive feedback expressed by teachers develops confidence and motivation over the long term (Wang, 2024). The trust in human aspects is additionally influenced by culture peculiarities that are common in various EFLs and are identified in systematic studies of GenAI in language classes (Zou et al., 2025). All of these discoveries add to the pre-AI feedback scholarship (Hyland and Hyland, 2006) and they place importance on the ability of GenAI to democratise access in resource scarce settings that also calls to the standardisation of approaches in SLA studies (Chapelle, 2001).\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eLimitations\u003c/h2\u003e \u003cp\u003eThis geographical sampling of East Asia, Middle East and Americas, underrepresenting all of Africa and Europe may reduce the generalizability of the findings and skew the insights on knowledge of cultures with stronger power distance, in which the teacher power is more promoted (Hofstede, 2001). Long-term effect assessment is confounded by the variability in the duration of intervention (weeks to semesters), which promotes short-term biases. Self-reported data is prone to social desirability, and methodological heterogeneity did not allow meta-analysis and instead required narrative synthesis. The non-native English situation with the AI technology may widen the digital gaps and discrimination issues when there is an ethical inefficiency or overtrust. The fast development of AI causes the risk of obsolescence (Escalante et al., 2023). These restrictions should be approached with reservations, and more searches and various sampling should be used in future studies.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eImplications and Future Directions\u003c/h2\u003e \u003cp\u003ePedagogically, These possibilities pedagogically propose the framework of hybrids in teaching EFL, which offers AI, preliminary error correction to reduce the workload of teachers in their tasks and preserving human resources to provide complex scaffolding (Asadi et al., 2025). This method can potentially scale in crowded classrooms, although it needs to take care of such issues as educator education in timely engineering as well as the reduction of AI biases so that it is a fair addition. Policymakers are supposed to put a priority on AI literacy programs in order to achieve equal access and avoid addiction.\u003c/p\u003e \u003cp\u003eFuture studies should use longitudinal designs to assess lasting effects, diversify sample by demographic and region and test AI tools of the next generation against developing teacher strategies. More sophisticated models, like the structural equation modeling of psychological constructs like motivation may enhance the knowledge regarding preference mechanisms (Aljasser, 2025). It is also crucial to carry out cross-cultural validations and ethical questions of the role of GenAI in the development of critical thinking (Abdi et al., 2025).\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis systematic review, synthesizing insights from 17 studies spanning 2023\u0026ndash;2025, illuminates a strategic division of labor in EFL writing feedback: generative AI is more suitable in reducing anxiety and surface-level correction, and teacher intervention is more suitable in higher-order writing skills, such as argumentation and coherence. Combining the effectiveness of AI with the human expertise in the form of the hybrid models appears to be the best solution, as it improves the quality of writing, satisfaction of the learners, and mental performance. Practically, the educators are expected to put in place the tiered methodology of AI in performing first mechanical corrections and teachers in performing contextual corrections with integrative engineering and bias audit mechanisms in place to make equity a reality in various EFL situations. This instructional system facilitates learning and avoids excessive dependence on technology. Supported by the recent evidence, the given review recommends consideration of technology-based pedagogies to empower worldwide EFL learners in dynamic learning settings. It encourages teachers and decision makers to consider the ethical use of AI by integrating AI into the education system to close gaps in teaching writing across the globe.\u003c/p\u003e "},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eAuthor Contributions All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by Y. Zhao and D. Li, who independently screened 2,623 records, extracted data based on the PICOS framework, and assessed methodological quality using MMAT 2018. G. The first draft of the manuscript was written by Y. Zhao. All authors commented on previous versions of the manuscript, refined the thematic synthesis, and read and approved the final manuscript.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eAll data and materials supporting the findings of this study are available in the Supplementary Materials accompanying this manuscript. These include the PRISMA 2020 checklist and flow diagram, full database search strategies, screening logs with inclusion and exclusion criteria, extracted records, data extraction forms, and synthesis tables. The materials are provided to ensure transparency and reproducibility of the systematic review process. This article does not contain any studies with human participants performed by any of the authors\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAbdi Tabari M, Kushki A, Wang Y (2025) Comparing the effects of teacher- and AI-mediated corrective feedback on accuracy, complexity, and quality in L2 written narratives. Comput Assist Lang Learn 1\u0026ndash;24. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/09588221.2025.2561608\u003c/span\u003e\u003cspan address=\"10.1080/09588221.2025.2561608\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAljasser A (2025) Investigating EFL students' perceptions of feedback: a comparative study of instructor and ChatGPT-generated responses in academic writing. \u003cem\u003eEducation and Information Technologies\u003c/em\u003e. Advance online publication. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10639-025-13649-8\u003c/span\u003e\u003cspan address=\"10.1007/s10639-025-13649-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlnemrat A, Aldamen H, Almashour M, Al-Deaibes M, AlSharefeen R (2025) AI vs. teacher feedback on EFL argumentative writing: a quantitative study. \u003cem\u003eFrontiers in Education\u003c/em\u003e, \u003cem\u003e10\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/feduc.2025.1614673\u003c/span\u003e\u003cspan address=\"10.3389/feduc.2025.1614673\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAsadi M, Ebadi S, Mohammadi L (2025) The impact of integrating ChatGPT with teachers\u0026rsquo; feedback on EFL writing skills. Think Skills Creativity 56:101766. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.tsc.2025.101766\u003c/span\u003e\u003cspan address=\"10.1016/j.tsc.2025.101766\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarrot JS (2023) Using ChatGPT for second language writing: Pitfalls and potentials. Assess Writ 57:100745. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.asw.2023.100745\u003c/span\u003e\u003cspan address=\"10.1016/j.asw.2023.100745\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBitchener J, Ferris DR (2012) Written Corrective Feedback in Second Language Acquisition and Writing. Routledge. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4324/9780203832400\u003c/span\u003e\u003cspan address=\"10.4324/9780203832400\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCao S, Zhong L (2023) \u003cem\u003eExploring the effectiveness of ChatGPT-based feedback compared with teacher feedback and self-feedback: Evidence from Chinese to English translation\u003c/em\u003e [Preprint]. arXiv. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.48550/arXiv.2309.01645\u003c/span\u003e\u003cspan address=\"10.48550/arXiv.2309.01645\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChapelle CA (2001) Computer Applications in Second Language Acquisition. Cambridge University Press, Cambridge\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChoi W (2025) Exploring high school students' preferences for English writing feedback: Non-native English teacher vs. native English teacher vs. ChatGPT. SAGE Open 15(4):1\u0026ndash;16. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/21582440251383693\u003c/span\u003e\u003cspan address=\"10.1177/21582440251383693\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCobb T (2008) Commentary: Response to mcquillan and krashen (2008). \u003cem\u003eLanguage Learning \u0026amp; Technology\u003c/em\u003e, \u003cem\u003e12\u003c/em\u003e(1), 109\u0026ndash;114. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.64152/10125/44134\u003c/span\u003e\u003cspan address=\"10.64152/10125/44134\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCrowe M, Sheppard L (2011) A review of critical appraisal tools show they lack rigor: Alternative tool structure is proposed. J Clin Epidemiol 64(1):79\u0026ndash;89\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDerakhshan A (2025) EFL students\u0026rsquo; perceptions about the role of generative artificial intelligence (GAI)-mediated instruction in their emotional engagement and goal orientation: A motivational climate theory (MCT) perspective in focus. Learn Motiv 90:102114. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.lmot.2025.102114\u003c/span\u003e\u003cspan address=\"10.1016/j.lmot.2025.102114\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEscalante J, Pack A, Barrett A (2023) AI-generated feedback on writing: Insights into efficacy and ENL student preference. Int J Educational Technol High Educ 20(1):57. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s41239-023-00425-2\u003c/span\u003e\u003cspan address=\"10.1186/s41239-023-00425-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFerris D, Hyland K, Hyland F (2006) Does error feedback help student writers? New evidence on the short- and long-term effects of written error correction. Feedback in Second Language Writing: Contexts and Issues. Cambridge University Press, Cambridge, pp 81\u0026ndash;104\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHenderson M, Bearman M, Chung J, Fawns T, Shum SB, Matthews KE, De Mello Heredia J (2025) Comparing Generative AI and teacher feedback: student perceptions of usefulness and trustworthiness. Assess Evaluation High Educ 1\u0026ndash;16. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/02602938.2025.2502582\u003c/span\u003e\u003cspan address=\"10.1080/02602938.2025.2502582\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHofstede G (2001) Culture's consequences: Comparing values, behaviors, institutions and organizations across nations. Sage\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHyland K, Hyland F (2006) Feedback on second language students\u0026rsquo; writing. Lang Teach 39(2):83\u0026ndash;101. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1017/S0261444806003399\u003c/span\u003e\u003cspan address=\"10.1017/S0261444806003399\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLo N, Chan S, Wong A (2025) Evaluating teacher, AI, and hybrid feedback in English Language Learning: Impact on student motivation, quality, and performance in Hong Kong. SAGE Open 15(3). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/21582440251352907\u003c/span\u003e\u003cspan address=\"10.1177/21582440251352907\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMakwana V (2025) A Comparative Analysis of AI-Powered and Teacher-Led Feedback: Investigating Student Perceptions and Writing Performance. J Engl Lang Teach 67(1):3\u0026ndash;12. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://journals.eltai.in/jelt/article/view/JELT670102\u003c/span\u003e\u003cspan address=\"https://journals.eltai.in/jelt/article/view/JELT670102\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMekheimer M (2025) Generative AI-assisted feedback and EFL writing: A study on proficiency, revision frequency and writing quality. Discover Educ 4(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003eArticle 170. https://doi.org/10.1007/s44217-025-00602-7\u003c/span\u003e\u003cspan address=\"Article 170. 10.1007/s44217-025-00602-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMu\u0026ntilde;oz Mu\u0026ntilde;oz BC, Nassaji H, Bello Carrillo FI (2025) ChatGPT-generated versus human direct corrective feedback on L2 writing. System 134:103805. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.system.2025.103805\u003c/span\u003e\u003cspan address=\"10.1016/j.system.2025.103805\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePage MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, Shamseer L, Tetzlaff JM, Akl EA, Brennan SE, Chou R, Glanville J, Grimshaw JM, Hr\u0026oacute;bjartsson A, Lalu MM, Li T, Loder EW, Mayo-Wilson E, McDonald S, Moher D (2021) The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 372:n71. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/bmj.n71\u003c/span\u003e\u003cspan address=\"10.1136/bmj.n71\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePluye P, Hong QN (2014) Combining the power of stories and the power of numbers: Mixed Methods Research and Mixed Studies Reviews. Annu Rev Public Health 35:29\u0026ndash;45\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRamazani A, Bijani H, Oroji MR (2025) Comparative analysis of AI vs. human feedback effects on IELTS candidates' writing performance. J Foreign Lang Teach Translation Stud 10(1):17\u0026ndash;40. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.22034/efl.2025.493559.1334\u003c/span\u003e\u003cspan address=\"10.22034/efl.2025.493559.1334\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRyan RM, Deci EL (2020) Intrinsic and extrinsic motivation from a self-determination theory perspective: Definitions, theory, practices, and future directions. Contemp Educ Psychol 61:101860. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.cedpsych.2020.101860\u003c/span\u003e\u003cspan address=\"10.1016/j.cedpsych.2020.101860\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShen L, Wang S, Xin Y (2025) EFL students\u0026rsquo; writing engagement and AI attitude in GenAI-assisted contexts: A mixed-methods study grounded in SDT and TAM. Learn Motiv 92:102168. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.lmot.2025.102168\u003c/span\u003e\u003cspan address=\"10.1016/j.lmot.2025.102168\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSoloveyo OZ (2024) Comparing peer, ChatGPT, and teacher corrective feedback in EFL writing: Students' perceptions and preferences. Technol Lang Teach Learn 6(3):1\u0026ndash;23. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.29140/tltl.v6n3.1482\u003c/span\u003e\u003cspan address=\"10.29140/tltl.v6n3.1482\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSoori A, Khojasteh L, Javed F (2025) Comparing teacher e-feedback, AI feedback, and hybrid feedback in enhancing EFL writing skills. Technol Lang Teach Learn 7(3). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003eArticle 102626. https://doi.org/10.29140/tltl.v7n3.102626\u003c/span\u003e\u003cspan address=\"Article 102626. 10.29140/tltl.v7n3.102626\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSweller J (1988) Cognitive load during problem solving: Effects on learning. Cogn Sci 12(2):257\u0026ndash;285. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1207/s15516709cog1202_4\u003c/span\u003e\u003cspan address=\"10.1207/s15516709cog1202_4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTopuz AC, Yıldız M, Taşlıbeyaz E, Polat H, Kurşun E (2025) Is generative AI ready to replace human raters in scoring EFL writing? Comparison of human and automated essay evaluation. Educational Technol Soc 28(3):36\u0026ndash;50. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.30191/ETS.202507_28(3).SP04\u003c/span\u003e\u003cspan address=\"10.30191/ETS.202507_28(3).SP04\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVygotsky LS (1978) Mind in society: The development of higher psychological processes. Harvard University Press. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.hup.harvard.edu/catalog.php?isbn=9780674576292\u003c/span\u003e\u003cspan address=\"https://www.hup.harvard.edu/catalog.php?isbn=9780674576292\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang D (2024) Teacher- versus AI-generated (Poe application) corrective feedback and language learners' writing anxiety, complexity, fluency, and accuracy. Int Rev Res Open Distrib Learn 25(3):37\u0026ndash;56. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.19173/irrodl.v25i3.7646\u003c/span\u003e\u003cspan address=\"10.19173/irrodl.v25i3.7646\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWarschauer M, Tseng W, Yim S, Webster T, Jacob S, Du Q, Tate T (2023) The affordances and contradictions of AI-generated text for writers of english as a second or foreign language. J Second Lang Writ 62:101071. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jslw.2023.101071\u003c/span\u003e\u003cspan address=\"10.1016/j.jslw.2023.101071\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang Y, Yao S (2025) Beyond linguistic proficiency: emotional, cognitive, and cultural pathways of Generative AI in EFL Well-Being\u0026mdash;A Technology‐Humanities‐Culture Synthesis. Int J Appl Linguistics. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/ijal.12832\u003c/span\u003e\u003cspan address=\"10.1111/ijal.12832\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu H, Xie Q (2025) Generative AI vs. teachers: Feedback quality, feedback uptake, and revision. Lang Teach Res Q 47:113\u0026ndash;137. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.32038/ltrq.2025.47.07\u003c/span\u003e\u003cspan address=\"10.32038/ltrq.2025.47.07\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Z (2020) Engaging with automated writing evaluation (AWE) feedback on L2 writing: Student perceptions and revisions. Assess Writ 43:100439. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.asw.2019.100439\u003c/span\u003e\u003cspan address=\"10.1016/j.asw.2019.100439\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Z, Aubrey S, Huang X, Chiu TKF (2025) The role of generative AI and hybrid feedback in improving L2 writing skills: a comparative study. Innov Lang Learn Teach 1\u0026ndash;19. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/17501229.2025.2503890\u003c/span\u003e\u003cspan address=\"10.1080/17501229.2025.2503890\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZou S, Guo K, Wang J, Liu Y (2025) Investigating students\u0026rsquo; uptake of teacher- and ChatGPT-generated feedback in EFL writing: a comparison study. Comput Assist Lang Learn 1\u0026ndash;30. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/09588221.2024.2447279\u003c/span\u003e\u003cspan address=\"10.1080/09588221.2024.2447279\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"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":"humanities-and-social-sciences-communications","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"palcomms","sideBox":"Learn more about [Humanities \u0026 Social Sciences Communications](http://www.nature.com/palcomms/)","snPcode":"41599","submissionUrl":"https://submission.springernature.com/new-submission/41599/3","title":"Humanities and Social Sciences Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Generative AI, EFL writing feedback, Hybrid models, Learner psychology","lastPublishedDoi":"10.21203/rs.3.rs-8386846/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8386846/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe introduction of GenAI into instruction in a second language puts the conventional focus on teacher feedback in question. Although tools such as ChatGPT are quickly spreading, whether algorithmic or human feedback are more effective in education is an issue in pedagogical discussion. The systematic review consolidates findings on 17 empirical studies (2023\u0026ndash;2025) with a unique focus on GenAI and teacher feedback comparison in EFL writing situation. With reference to the self-determination and cognitive load theories, our analysis reveals the steady division of labor in feedback efficacy where GenAI performs more effectively as a nurturing factor in reducing anxiety and surface-level error-correction skills than teacher feedback, which performs better in fostering higher-order level argumentation, coherent and metacognitive awareness. Most importantly, this review empirically confirms the superiority in hybrid feedback models, which allow using AI to achieve immediacy and accuracy, with a human-focused viewpoint used only to respond to the situation. We suggest that GenAI should exist and operate in a culturally competent, collaborative structure intelligence that does not replace, but adds the scaffold that will streamline the work of the teacher and increase student autonomy. The results can serve as an evidence-based guideline to applying AI in writing classes and reducing the risks of overdependence.\u003c/p\u003e","manuscriptTitle":"A Systematic Comparative Review of Generative AI vs. Teacher Feedback in EFL Writing: Learner Perceptions, Preferences, and Psychological Factors","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-27 18:14:22","doi":"10.21203/rs.3.rs-8386846/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-05-17T15:14:50+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"76059367316092334493084207601141530361","date":"2026-05-02T05:32:03+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"5689960031065068960984989369539333051","date":"2026-05-01T16:10:58+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-01T05:58:54+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"13892365314774710042984001876567884742","date":"2026-04-19T15:47:22+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-08T13:10:02+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"76059367316092334493084207601141530361","date":"2026-01-29T23:51:38+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"21781713914522873113491416763120631108","date":"2026-01-23T02:26:02+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-01-22T14:21:46+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-01-22T14:13:14+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-01-08T07:04:43+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-01-06T14:04:07+00:00","index":"","fulltext":""},{"type":"submitted","content":"Humanities and Social Sciences Communications","date":"2026-01-06T13:53:34+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"humanities-and-social-sciences-communications","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"palcomms","sideBox":"Learn more about [Humanities \u0026 Social Sciences Communications](http://www.nature.com/palcomms/)","snPcode":"41599","submissionUrl":"https://submission.springernature.com/new-submission/41599/3","title":"Humanities and Social Sciences Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"3870a87e-211f-467f-97b2-19c6359aed7b","owner":[],"postedDate":"January 27th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"","date":"2026-05-17T15:14:50+00:00","index":77,"fulltext":""},{"type":"reviewerAgreed","content":"76059367316092334493084207601141530361","date":"2026-05-02T05:32:03+00:00","index":76,"fulltext":""},{"type":"reviewerAgreed","content":"5689960031065068960984989369539333051","date":"2026-05-01T16:10:58+00:00","index":75,"fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-01T05:58:54+00:00","index":65,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":61681548,"name":"Physical sciences/Mathematics and computing"},{"id":61681549,"name":"Biological sciences/Psychology"},{"id":61681550,"name":"Social science/Psychology"}],"tags":[],"updatedAt":"2026-01-27T18:14:22+00:00","versionOfRecord":[],"versionCreatedAt":"2026-01-27 18:14:22","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8386846","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8386846","identity":"rs-8386846","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}