Author contributions Establish Clear Guidelines: Educational institutions should create clear

The gap

Evidence profile

Sourced from the future work and recommendations of the source papers, classified as general, drawn from work published between 2025 and 2026, spanning 4 journals. Those papers have been cited 9 times in total.

Supporting evidence — 5 representative gaps

• The Influence of Generative AI Usage Styles on Creative Self-Beliefs: Findings from a Longitudinal Study in Design Education (2026) · doi

  The findings of this study highlight the limitations of understanding the educational effects of generative AI solely in terms of its use or non-use, and suggest that its impact depends on usage style and the nature of the human–AI relationship. Although generative AI was primarily perceived as a tool for improving task efficiency, usage styles such as feedback provision and meaning summarization were associated with increases in CSE. However, this study has several limitations. First, the analysis was based on a single course context, and the sample was limited to 64 participants who completed all measurements; therefore, caution is required when generalizing the findings. Second, in the analysis by AI usage style, some categories included only a small number of participants, and thus, the results should be considered exploratory. Finally, the study was relatively short, and it is possible that the relationship between users and AI had not yet been sufficiently established. Future research should adopt a longer-term longitudinal approach to examine how interactions with generative AI evolve over time and how these changes influence CSB. In addition, incorporating factors such as perceived control and sense of agency toward AI into measurement frameworks may enable a more nuanced understanding of how human–AI relationships shape creativity. CRediT authorship contribution statement Riku Okamoto: Conceptualization, Methodology, Investigation, Formal analysis, Visualization, Writing – original draft, Writing – review & editing. Akiyoshi Inasaka: Conceptualization, Supervision, Resources, Investigation, Writing – review & editing.

  generalfuture work
  Keywords: generative usage writing limitations understanding style human relationship perceived participants conceptualization investigation review editing highlight
• Artificial Intelligence-Assisted Creative Writing in Malay Literature Education: Students’ Perceptions and a Human-AI Collaborative Learning Model (2026) · Indonesian Journal on Learning and Advanced Education (IJOLAE) · doi

  collaborative model proposed in this study, where AI functions as a facilitator while human authorship remains central to literary creation. reinforces the c. Solutions In this study, respondents offered several suggestions to address the challenges they identified. As mentioned in Figure 1, these suggestions were grouped into four themes: balancing usage, ideation, experimenting with literary work, and improvisation and analysis of work. These suggestions show that students were not rejecting AI, but were proposing a responsible and balanced model of AI use in literary learning. 1) Balancing Usage In response to the challenges identified, respondents proposed several strategies for the responsible integration of AI in literary education. The most common recommenda- tion was maintaining a balanced approach to AI usage. Students emphasised that AI should complement, rather than replace, human crea- tivity and critical thinking throughout the wri- ting process. “…My suggestion is that AI practitioners and us- ers also need to have balance in the production of works, such as creating their own literary works but delivering them through the use of AI as an option, and not producing works entirely using AI...” (Respondent 2) “My suggestion is that when wanting to produce a literary work, use AI only as a reference and then use one’s own ideas and creativity so that the literary work produced is of higher quality…” (Respondent 6) “Taking into account and balancing what is given by AI and being wise in making use of it so that it can give an individual a broad perspective. How- ever, do not take one hundred per cent of what AI gives and never use AI without balancing the ele- ments that already exist in something. For exam- ple, even though AI gives an example of a pantun stanza, we need to balance whether the stanza given by AI corresponds with the values of pan- tun...” (Respondent 10) “In my opinion, AI can still be used as long as it is in moderation and not relied upon one hundred per cent. It may be used when making a mind map of the ideas that are to be used in the work.” (Re- spondent 21) Indonesian Journal on Learning and Advanced Education (IJOLAE)| p-ISSN 2655-920x, e-ISSN 2656-2804 Vol. 8 (2) (2026) 392-416 409 Artificial Intelligence-Assisted Creative Writing in Malay Literature Education: Students’ Perceptions and a Human–AI Collaborative Learning Model First, the respondents consistently empha- sised moderation in the creative process by balancing the usage of AI in creating creative literary work. Respondent 2 suggests that writers should first create their own literary works and then use AI as an optional tool for delivery or enhancement, rather than allowing AI to generate the entire piece. Similarly, Re- spondent 6 proposes that AI should be used only as a reference, while the main ideas and creativity must come from the writer to ensure higher quality and more authentic work. Re- spondent 10 further stresses the need to criti- cally evaluate and balance what AI provides, warning against taking AI output entirely without considering existing literary elements and values. For instance, when AI suggests a ‘pantun’ stanza, students should still check whether it truly fits the local culture and values. Re- spondent 21 reinforces the idea of modera- tion, suggesting that AI can be used for spe- cific tasks such as creating mind maps of ideas, but should not become a complete sub- stitute for human effort. These responses show that students advocate a collaborative model, where AI is a supportive tool while hu- man judgment, creativity, and cultural under- standing remain central in the production of Malay literary works. This balanced model is the main theoretical contribution of the fin- dings, as it frames responsible AI use in Ma- lay Literature as collaboration between AI as- sistance and human-cultural authorship.

  generalfuture work
  Keywords: literary model human balancing students works used usage respondent ideas spondent collaborative respondents suggestions responsible
• Artificial Intelligence in EFL Writing Instruction: A Critical Narrative Review of Feedback, Pedagogical Integration, and Learning Outcomes (2026) · Fundamental Scientific Reports in Multidisciplinary Areas · doi

  Future research should examine what learners can do after AI support is removed. A revised draft may improve immediately after ChatGPT or an AWE tool is used, but this does not show whether the learner has developed independent control. Longitudinal studies should therefore compare supported revision with later unaided writing. Measures could include accuracy, organization, lexical choice, self-editing behavior, and students' explanations of their own revision decisions. More work is also needed in secondary-school EFL classrooms. Much of the available evidence comes from higher education, where students may have stronger academic literacy and more developed self-regulation. Younger learners may need clearer boundaries, more modeling, and more teacher supervision. Studies should examine how age, proficiency, task type, teacher guidance, and assessment context interact instead of treating AI use as a single variable. Finally, feedback literacy and AI literacy should be treated as learning outcomes. Students need to learn how to ask useful questions, compare feedback sources, identify unsuitable suggestions, preserve their own meaning, and disclose assistance honestly. Teachers need support as well, since designing AI-mediated writing tasks requires time, policy clarity, and professional judgment. The most important future question is not only whether AI improves scores, but whether it helps learners become more capable and responsible writers.

  generalfuture work
  Keywords: learners whether students literacy need future examine support developed compare revision writing self teacher feedback
• AI-Mediated Writing Instruction in Higher Education: A Systematic Review of Empirical Evidence (2026) · Journal of Education and Training Studies · doi

  Building on these limitations, several directions for future research emerge. First, there is a clear need for longitudinal and developmental studies that examine how sustained engagement with generative AI tools influences writing competence over time. Such research should move beyond surface-level textual improvements to investigate the evolution of argumentation, rhetorical awareness, and learner autonomy across academic trajectories. Second, future studies should prioritize theoretical integration and construct clarity. Researchers are encouraged to articulate explicit pedagogical and learning theories—such as socio-constructivism, self-regulated learning, or distributed cognition—when designing AI-mediated writing interventions. Clear alignment between instructional design, theoretical framework, and outcome measures will strengthen both internal validity and cumulative knowledge-building. Third, greater attention should be paid to instructional design variables, including levels of scaffolding, degrees of instructor mediation, and the explicit teaching of AI literacy. Comparative studies examining guided versus unguided AI use, or hybrid versus AI-dominant instructional models, would offer valuable insight into pedagogical conditions that support deep learning rather than dependency. Fourth, future research must more systematically address ethical and assessment-related questions. Empirical investigations into authorship attribution, transparency practices, process-based assessment models, and student ethical 134 Journal of Education and Training Studies Vol. 14, No. 4; October 2026 reasoning are essential for developing pedagogically defensible responses to generative AI. Such work would help shift the discourse from policy-driven restriction toward evidence-informed instructional innovation. Finally, expanding the scope of research to include diverse educational contexts—including graduate writing, disciplinary writing, multilingual classrooms, and underrepresented regions—will be critical for advancing a globally relevant understanding of AI-mediated writing instruction. Cross-cultural and cross-institutional comparative studies, in particular, hold promise for identifying context-sensitive pedagogical principles.

  generalfuture work
  Keywords: writing instructional future pedagogical learning building clear generative theoretical explicit mediated design including comparative versus
• The impact of ChatGPT on academic integrity in medical education: a developing nation perspective (2025) · Frontiers in Education · cited 9× · doi

  Author contributions Establish Clear Guidelines: Educational institutions should create clear, internationally standardized policies outlining the permissible use of AI tools like ChatGPT. These guidelines should be binding and enforceable across institutions. • Redefine Authorship and Compliance: Authorship standards for publications including legal implications for non-compliance. Institutions should also set limits on AI usage for manuscript writing. should be redefined, • Integrate AI into Learning Sessions: Incorporate AI tools into student learning activities, such as small group discussions and project-based learning. Redesign teaching sessions to promote the effective and ethical use of AI technology. AJ: Conceptualization, Visualization, Writing – original draft, Writing – review & editing. RA: Data curation, Formal analysis, Writing – original draft, Writing – review & editing. GF: Data curation, Formal analysis, Writing – original draft, Writing – review & editing. MZB: Resources, Supervision, Validation, Writing – original draft, Writing – review & editing.

  generalrecommendations
  Keywords: writing original draft review editing institutions learning clear guidelines tools rede authorship compliance sessions curation

Questions about this gap