Open research questions in Online Learning and Analytics

What the literature leaves open

• The evidence therefore supports a cautiously positive conclusion: AI can enhance educational administration when it is embedded in redesigned workflows, staff development, human validation, and risk-based governance, but the current evidence base is insufficient for strong causal or universal claims.

  Artificial Intelligence in Educational Administration: A Systematic Evidence Review and Exploratory Meta-Analysis of Empirical Studies, 2020–2025 · 2026 · DOI

• Future research should focus on empirical validation of the HAGD Framework, longitudinal and cross-cultural investigations, and the examination of governance mechanisms, fairness, and hybrid human-AI assessment models to ensure that AI- http://ijlter.

  AI-driven Assessment in Education: A Bibliometric Analysis and Framework for Assessment Design and Governance · 2026 · DOI

• Schools should be given guidance and support in developing clear policies and providing training for the appropriate use of AI tools in English assignments. The emphasis should be on AI tools as only one of the means or sources of help, and not as the ones replacing students’ work done independently. Some of the ways teachers can introduce AI in writing assignments include allowing students to use these tools to a limited extent in writing, and there can also be other types of work that will require students to think on their own, do a critical analysis, and solve problems, for example, keeping reflection journals, doing peer reviews, and writing creatively without using AI. Additionally, students’ writing skills and academic integrity can be enhanced through regular monitoring and feedback, even as they get help from AI in learning grammar, vocabulary, and working more efficiently. Furthermore, practical sessions can be held to familiarize students with the appropriate use of AI for such purposes as research, summarization, and idea generation. Yet, what is crucial is that students’ works remain their own and as clear and different from others as possible. https://journals.stecab.comStecab PublishingPage Journal of ICT and Education (JICTE), 1(2), 39-51, 2026 48 REFERENCES Abbas, N., Ali, I., Manzoor, R., Hussain, T., & Hussaini, M. H. A. (2023). Role of artificial intelligence tools in enhancing students’ educational performance at higher levels. Journal of Artificial Intelligence, Machine Learning and Neural Network, 3(05), 36-49. https://doi.org/10.55529/ jaimlnn.35.36.49. Akgun, S., & Greenhow, C. (2022). Artificial Intelligence (AI) in Education: Addressing Societal and Ethical Challenges in K-12 Settings. Proceedings of International Conference of the Learning Sciences, ICLS. Alharbi, M. (2022). The impact of artificial intelligence on academic writing: A case study of EFL students. Journal of Language and Education, 8(3), 45-58. Annamalai, N. (2025). Factors affecting English language high school teachers switching intention to ChatGPT: A Push- Pull-Mooring theory perspective.

  Senior High School Students’ and English Teachers’ Perceptions of AI Tools in English‑Related Tasks: Benefits and Concerns in Baguio City · 2026 · DOI

• While this enabled AI use for English learning purposes, future research could explore similar constructs with EFL learners at different educational levels, such as middle school students.

  Generative Artificial Intelligence Literacy Training for 12th Grade English as a Foreign Language Students: Effects on Artificial Intelligence Literacy, ChatGPT Usage, and Autonomous English Learning · 2026 · DOI

• Drawing on the study’s findings and their interpretation through the Technological Pedagogical Content Knowledge (TPACK) framework and Critical Digital Pedagogy (CDP), several recommendations emerge to strengthen the responsible and effective integration of artificial intelligence (AI) technologies in higher education, particularly within rural institutional contexts. First, higher education institutions should prioritise structured professional development programmes that equip lecturers with integrated technological, pedagogical, and disciplinary competencies required for meaningful AI-supported teaching. Such training should move beyond basic technological orientation to focus on pedagogical strategies for embedding AI tools within curriculum design, assessment practices, and student engagement processes, thereby strengthening lecturers’ TPACK capabilities. Second, universities should develop clear institutional policies and ethical guidelines governing the use of AI technologies in teaching, learning, and assessment. These frameworks should provide explicit guidance on issues such as academic integrity, responsible AI use, authorship, and transparency in AI-assisted learning to reduce uncertainty among both lecturers and students. Third, institutional leadership should prioritise investments in digital infrastructure, including reliable internet connectivity, updated digital learning platforms, and accessible computer facilities, particularly for students located in rural or resource-constrained environments. Addressing these structural inequalities is essential for ensuring that AI technologies do not reproduce existing educational disparities. Fourth, lecturers should be encouraged to integrate AI into teaching in ways that promote critical digital literacy, enabling students to evaluate, question, and responsibly engage with AI-generated knowledge rather than relying on it uncritically. This approach aligns with the principles of Critical Digital Pedagogy by positioning students as active and reflective participants in digitally mediated learning environments. Finally, future research should further explore the long-term pedagogical implications of AI adoption across diverse higher education contexts, with a particular focus on how lecturers develop sustained technological pedagogical content knowledge and how institutions can design inclusive digital strategies to ensure equitable access to emerging educational technologies. Collectively, these recommendations underscore that the transformative potential of AI in higher education can only be realised through coordinated institutional support, pedagogical innovation, and critical engagement with the social and ethical dimensions of digital technologies.

  Artificial intelligence in education: A phenomenological study of opportunities, ethical tensions, and digital inequality in South African universities · 2026 · DOI

• Artificial intelligence (AI) is increasingly reshaping higher education, yet the psychological mechanisms that determine when and for whom AI-assisted learning translates into academic success remain insufficiently understood, particularly within developing educational contexts.

  The moderating effect of academic self-efficacy on artificial intelligence-assisted learning and academic achievement · 2026 · DOI

• SUGGESTIONS FOR FURTHER STUDIES To strengthen the evidence base and applicability of the IAPF, future research should focus on the following priority areas. Its effectiveness and practical applicability in real classroom settings, including its capacity to produce measurable improvements in pre-service teacher AI competency and pedagogical integration, remain to be established through future empirical investigation.

  Developing an Integrated AI Pedagogical Framework for Pre-Service Teachers · 2026 · DOI

• ADVANCED RESEARCH Future studies are recommended to explore teachers’ pedagogical adaptation to AI using longitudinal designs to capture changes over time.

  Teachers' Pedagogical Adaptation in the Use of Artificial Intelligence in Secondary School Learning · 2026 · DOI

• Volume 5 Issue 2 ISSN 2790-9441 T E S O L C o m m u n i c a t i o n s | 41 More broadly, the findings contribute new evidence to the limited research on AI use in high- stakes exam preparation and offer a context-specific perspective from Vietnam, where teacher- centered approaches remain dominant.

  Pioneering AI-Enhanced IELTS Task Delivery in Vietnam: A Transformative Model for Intermediate Learners · 2026 · DOI

• Importantly, this study extends the literature by indicating that these effects occur not only in general EFL settings, but also in exam-oriented, high-stakes preparation, an area that has been underexplored.

  Pioneering AI-Enhanced IELTS Task Delivery in Vietnam: A Transformative Model for Intermediate Learners · 2026 · DOI

• Future research could strengthen and extend the contributions of this study in several ways. Larger, multi-site studies involving diverse learner populations would help determine the broader applicability of AI-supported task delivery in high-stakes exam preparation. Longer- term interventions are needed to examine whether gains in receptive and productive skills are sustained over time. Research should also explore optimized feedback designs, including simplified or bilingual explanations, to address cognitive load challenges identified in this study. Additionally, studies incorporating motivational supports—such as gamification, adaptive reminders, or structured reflection tools—may help sustain engagement and learner autonomy across extended courses. Finally, future evaluations should systematically document the technical reliability of AI platforms to better understand its influence on learner trust, persistence, and outcomes.

  Pioneering AI-Enhanced IELTS Task Delivery in Vietnam: A Transformative Model for Intermediate Learners · 2026 · DOI

• 35 ERIES Journal volume 19 issue 1Printed ISSN 2336-2375Electronic ISSN 1803-1617 Furthermore, this study is limited to a single open-source dataset from a single higher education institution in Portugal, thereby limiting the generalizability of the findings.

  Unpacking the Black Box: A Hybrid XAI Framework for AutoGluon-Based Multiclass Student Outcome Prediction · 2026 · DOI

• Based on the findings of this study, there are some recommendations that can be made for teachers, inspectors, school headmasters, and education policymakers. First, teachers should believe in the wide range of benefits that AI tools could bring to the language class. They should take risks using these tools and encourage students to use them too. They should also take some free online courses about integrating AI in education. Second, inspectors can organize seminars, workshops, and study days to empower teachers with pedagogical guidelines on integrating AI into their lessons. Third, the headmaster should encourage AI use in schools and offer facilities and technical support for teachers. Finally, education policymakers should develop clear policies for AI use in education, integrate AI training into teacher training institutions, and equip students with a reliable internet connection and digital tools. Furthermore, researchers can conduct further research on AI integration in Moroccan classrooms.

  English Language Teachers’ Perceptions of Artificial Intelligence Use in the EFL Classroom · 2026 · DOI

• Based on the findings of the study, the following recommendations were made: i. Governments and educational institutions should prioritize reliable electricity and high-speed internet as fundamental prerequisites for effective AI integration in teaching and learning. Adequate infrastructure ensures that AI tools can move beyond peripheral tasks to support core instructional activities. ii. Educators need ongoing professional development that builds both technical skills and pedagogical expertise. Training should emphasize the meaningful application of AI in classrooms, focusing on adaptive teaching strategies, personalized learning, and innovative instructional practices. iii. AI tools should be adapted to the local sociocultural and technological environment rather than using a uniform approach. Context-sensitive integration increases relevance, acceptance, and sustainability, ensuring that AI applications address real educational needs. iv. Clear policies and guidelines should be established to protect student data, uphold academic integrity, and promote inclusivity. Ethical frameworks foster trust among teachers, students, and stakeholders, while guiding responsible AI adoption. v. AI should be introduced progressively, beginning with supportive functions and gradually moving into core pedagogical roles as teachers and students build competence. Collaboration between universities, government, and private sectors can help develop locally relevant solutions that align with global best practices while addressing specific challenges.

  A standardized framework for structured pedagogy and responsible knowledge (SPARK) Building for integrating artificial intelligence into 21st century classrooms · 2026 · DOI

• Overcoming this stagnation necessitates a concerted national effort that moves beyond mere technological acquisition. It requires a strategic commitment to: 1. Policy and Empowerment: Crafting clear, actionable policies that demystify the use of AI, while simultaneously investing in comprehensive training that empowers lecturers to become confident and critical users of these new tools. 2. Infrastructure and Equity: Committing to significant investment not only in institutional infrastructure but also in national initiatives aimed at closing the digital literacy and access gaps for all students. Ultimately, for South Africa, the path forward is not about simply "adopting AI." It is about thoughtfully and equitably weaving it into the educational fabric. Failing to address these core challenges of educator empowerment and student equity will ensure AI remains a source of exclusion, further cementing the nation's position on the wrong side of the global innovation divide. CONCLUSION The stagnant adoption of Artificial Intelligence in South Africa’s online teaching and learning environments results from a convergence of structural, institutional, and human-related challenges. This issue represents a dual crisis of confidence and access, both of which impede progress and collectively constrain the potential of AI in higher education. A significant crisis of confidence persists among academic staff and decision-makers. The absence of well-defined institutional policies, ethical standards, and governance frameworks has generated uncertainty regarding the appropriate use of AI in teaching and learning. Insufficient training and professional development leave many lecturers unprepared and hesitant to adopt AI tools. This uncertainty undermines pedagogical authority and intensifies concerns about academic integrity and the ethical use of generative AI. Without structured support and capacity-building initiatives, academic staff remain reluctant to integrate AI into their practices, which contributes substantially to stagnation across higher education institutions. Alongside this is a persistent crisis of access, shaped by South Africa’s longstanding socio-economic disparities. Many students continue to face significant barriers, including limited device availability, unreliable internet connections, and unaffordable data costs that hinder their ability to fully engage with AI- enabled learning. Despite institutional efforts to address these inequities, economic pressures and structural inequalities continue to impede meaningful progress. As long as these obstacles continue, the benefits of AI-enhanced education will remain out of reach for a substantial portion of the student population. The interdependence of these two crises perpetuates a cycle of stagnation.

  Overcoming Inertia: A systematic review of the stagnant integration of Artificial Intelligence in Online Learning at Universities of Technology in South Africa · 2026 · DOI

• Another research field that could be explored is how to use XAI along with new technologies such as adaptive learning platforms and virtual classrooms, to enhance personalized learning, whilst maintaining the level of trust and transparency.

  Integrating Explainable Artificial Intelligence into Education: Building Trustworthy and Transparent Intelligent Learning Systems · 2026 · DOI

• The literature base skews toward community colleges, leaving research universities, historically Black colleges and universities, tribal colleges, and international contexts underrepresented, while longitudinal studies tracking gains into advanced programs and post-graduation outcomes remain sparse.

  AI-Augmented Pedagogy in Higher Education · 2026 · DOI

• To this end, future research should explore how to structure teacher PD to be more meaningful and impactful, as well as investigate effective pedagogical approaches for teaching AI. However, achieving this goal requires further research, as our review included only 36 studies, which is insufficient for constructing a robust, stable, and generalizable AI literacy framework.

  A Systematic Review Mapping of AI Literacy Progression in K–12 · 2026 · DOI

• Further, future investigations could investigate differentiated instruction approaches based on learner profiles to further individualize the development of AI literacy. The findings are limited due to the small sample size and short duration of the intervention.

  Developing AI Literacy Through a Holistic Framework among Pakistani University Students · 2026 · DOI

• Several recommendations for research and practice can be developed based on these conclusions and limitations. Longitudinal studies should be designed to understand how educators and students integrate AI tools over time. By including data over a period of time, it will not only give a more comprehensive view of the immediate effects of the tools but will also take into account long-term uptake, resistance, and organizational change. More research will need to be completed on policy development based on the ethics of bias and established privacy concerns in the use of such tools, as well as issues relating to academic integrity. Understanding how organizations design and develop governance and accountability frameworks for responsible protest use is essential to the responsible use of AI in education. Qualitative approaches and methods, such as interviews and focus group studies, are also important in understanding the lived experiences of stakeholders, the local context of challenges, and the integration of AI into practice. From a practice perspective, organizations or institutions need to engage educators and systems in how to develop an AI literacy curriculum, so that educators can use AI tools responsibly and create structures to critically assess the outputs of AI content. Institutions must regularly audit AI tools and develop transparent processes for using AI in learning spaces in order to see risk assessment both in time and function in reducing bias and misuse, as well as reducing overreliance on these systems. Policymakers must recognize that the sustainability of these systems is not only about how technically efficient these educational technologies are, but also about how they can adapt and continue to evolve, in keeping with ethical standards, inclusivity, and academic integrity. A comprehensive strategy that integrates these dimensions is essential to ensure that AI adoption enhances rather than undermines the foundations of higher education.

  Transforming Language Teaching With AI: An Investigation of Opportunities, Challenges, and Ethical Imperatives · 2026 · DOI

• Can anyone suggest an AI tool to create multiple “Hello choice exams, but long ones, 250 questions? I need to make long exams based on specific content. I’ve tried ChatGPT but it’s not always very accurate…” Sharing an event, a video, and a tool were the three categories found in all three online lan- guage teacher communities (see Table 5). This seems to suggest a common emphasis on shar- ing practical resources and staying informed within the teaching community. The “asking a question” category received the highest number of comments, with an average of 5.67 com- ments per post. This seems to indicate that questions were not overlooked but rather attracted considerable traffic and responses.

  Navigating AI in language education: Exploring the role of online language teacher communities · 2026 · DOI

• The results, of course, should be examined in terms of the specific customs of teaching and assessment in the particular context of the study. For this reason, the patterns observed may not be applicable to the STEM fields, clinical and laboratory instructional disciplines, or professional programs where teaching and assessment customs are more different from the disciplines of the study.

  AI-assisted learning and the illusion of competence · 2026 · DOI

• Deep Learning Vocational education institutions are encouraged to adopt a more comprehensive and human-centered approach when introducing technologies into their teaching and learning systems. Leaders should invest in continuous and meaningful professional development that helps teachers build their digital skills while reducing feelings of pressure or technological stress. Training programs will be more effective if they emphasize hands-on practice, collaboration among peers, and mentoring support rather than relying solely on theoretical sessions. In addition, institutional policies should make the use of AI-based tools voluntary and linked to clear pedagogical goals, rather than enforcing them through administrative mandates.

  Student Acceptance of AI-Driven Deep Learning Tools: The Influence of Institutional Support and Teacher Competency in Vocational Schools · 2026 · DOI

• This study closes the diagnostic gap in educational analytics by developing and validating an XAIbased framework for early student risk detection. Conventional ML models in education have traditionally been black boxes that emphasize mathematical precision rather than pedagogical disclosure. With a predictive accuracy of 92.9% and a very high AUC-ROC of 0.977, when an optimized Random Forest model is used, the framework serves as a highly reliable early-warning system that reduces both false alarms and missed interventions. One of the main contributions of this work is the integration of XAI via SHAP analysis, which can convert a technical risk score into a personalized diagnostic roadmap by revealing the why behind a prediction, pointing to behavioral drivers such as absenteeism, past failures, and midterm grades. The system is not just a classification but an action. Importantly, a formal Fairness Audit is a crucial ethical element of the framework. Using the Demographic Parity Difference measure, the study found that the model is not biased and provides fair risk detection across gender groups. This guarantees that the academic assistance is grounded in objective behavioral patterns, with critical attention to algorithmic bias in automated educational tools. In the end, this research provides a viable, scalable solution for higher education institutions to transition from reactive to proactive student support. The suggested framework will ensure that AI is a collaborative pedagogical partner that empowers educators and learners with clear, evidence-based feedback. The system creates a setting in which students can self-reflect and act intelligently in their academic path by identifying specific behavioral obstacles. Future studies will focus on longitudinal classroom studies to empirically determine how these transparent diagnostic prompts can influence student retention and behavior change across a wide variety of institutional settings. DATA AVAILABILITY The dataset analyzed during the current study is publicly available in the UCI Machine Learning Repository: https://archive.ics.uci.edu/dataset/320/student+performance REFERENCES Aldowah, H., Al-Samarraie, H., & Fauzy, W. M. (2019). Educational data mining and learning analytics for 21st century higher education: A review and synthesis. Telematics and Informatics, 37, 13–49. https://doi.org/10.1016/j.tele.2019.01.007 17 AI-Driven Risk Detection for Metacognitive Growth Algarni, A., Abdullah, M., Allahiq, H., & Qahmash, A. (2023). Predicting at-risk students in higher education. International Journal of Intelligent Systems and Applications in Engineering, 11(3), 1229–1239. https://ijisae.org/index.php/IJISAE/article/view/3382 Al-Shabandar, R., Hussain, A. J., Liatsis, P., & Keight, R. (2019). Detecting at-risk students with early interventions using machine learning techniques. IEEE Access, 7, 149464–149478.

  Leveraging Explainable AI to Enhance Student Metacognition Through Early Risk Detection · 2026 · DOI

• formal Fairness Audit was performed to guarantee that risks were fairly detected across gender groups. The research introduced a novel human-focused framework that fills the gap between predictive analytics and pedagogical theory. It shows how XAI can help turn a technical risk score into a metacognitive prompt, encouraging data- driven conversations between educators and students and offering a clear roadmap for institutional interventions. The analysis revealed that the Random Forest model achieved 92.9% accuracy and an AUC-ROC of 0.977. SHAP analysis found school absenteeism and mid- term grades as the most important risk predictors. Individualized diagnostics (waterfall plots) in the system provided the necessary evidence through student self-reflection, and the audit of fairness ensured that the model supports gender groups equally. Educational institutions must implement risk prediction systems based on XAI to go beyond mere warning systems. Practitioners should employ individual- level diagnostics to tailor mentoring and motivate students to reflect on their learning behaviors through evidence-based self-reflection. Further studies are expected to include longitudinal pilot studies that quantify the actual behavioral effects of XAI-based prompts on student outcomes. Re- searchers are also advised to test the framework using larger, multi-institutional datasets to increase its generalizability. Transparent and fair ML systems can improve student retention and graduation rates, leading to better resource allocation and a more inclusive educational en- vironment. By focusing on student agency, these systems foster a more success- ful, self-aware workforce that benefits society in the long term. To establish the global applicability and ethical soundness of the XAI frame- work, future studies must explore real-time application of XAI to learning man- agement systems and test the cross-cultural validity of behavioral predictors.

  Leveraging Explainable AI to Enhance Student Metacognition Through Early Risk Detection · 2026 · DOI

Most-cited papers in Online Learning and Analytics

• Students’ voices on generative AI: perceptions, benefits, and challenges in higher education · International Journal of Educational Technology in Higher Education · 2023 · 1,603 citations
• Artificial intelligence in higher education: the state of the field · International Journal of Educational Technology in Higher Education · 2023 · 1,240 citations
• A comprehensive AI policy education framework for university teaching and learning · International Journal of Educational Technology in Higher Education · 2023 · 1,004 citations
• The effects of over-reliance on AI dialogue systems on students' cognitive abilities: a systematic review · Smart Learning Environments · 2024 · 977 citations
• Ethical principles for artificial intelligence in education · Education and Information Technologies · 2022 · 879 citations
• Ethics of AI in Education: Towards a Community-Wide Framework · International Journal of Artificial Intelligence in Education · 2021 · 828 citations
• Shaping the Future of Education: Exploring the Potential and Consequences of AI and ChatGPT in Educational Settings · Education Sciences · 2023 · 752 citations
• Adaptive Learning Using Artificial Intelligence in e-Learning: A Literature Review · Education Sciences · 2023 · 733 citations
• Towards Intelligent-TPACK: An empirical study on teachers’ professional knowledge to ethically integrate artificial intelligence (AI)-based tools into education · Computers in Human Behavior · 2022 · 697 citations
• Challenges and Opportunities of Generative AI for Higher Education as Explained by ChatGPT · Education Sciences · 2023 · 672 citations

Most recent work

• Artificial intelligence in higher education: a systematic review of its impact on student engagement and the mediating role of teaching methods · Frontiers in Education · 2026
• Six Institutional Intervention Areas to Support Ethical and Effective Student Use of Generative AI in Higher Education: A Narrative Review · Education Sciences · 2026
• Unpacking EFL teachers’ perceived opportunities and challenges of multimodal AI-driven language testing · Language Testing in Asia · 2026
• Exploring the Role of AI Technology Adoption in at-Risk Students' Absenteeism and Boredom · International Journal of TESOL Studies · 2026
• A systematic review of gender differences in students’ use of AI tools for learning in higher education · Discover Education · 2026
• Pedagogical partnerships with generative AI in higher education: how dual cognitive pathways paradoxically enable transformative learning · International Journal of Educational Technology in Higher Education · 2026
• Predicting student performance: A comprehensive review of machine learning, deep learning, and explainable AI approaches · Computers and Education: Artificial Intelligence · 2026
• Teacher education for artificial intelligence literacy through a self-determination theory perspective · European Journal of Teacher Education · 2026
• ARTIFICIAL INTELLIGENCE IN HIGHER EDUCATION: SUPPORTING SELF-DIRECTED LEARNING AND STUDENT AUTONOMY · Turkish Online Journal of Distance Education · 2026
• Teachers’ artificial intelligence (AI) literacy: an exploratory study · Smart Learning Environments · 2026

Related topics in Computer Science