Research Gaps in AI in Education: Opportunities for Malaysian Postgraduate Students
What Are the Biggest Research Gaps in AI in Education Right Now?
The biggest research gaps in AI in education include the lack of explainability in AI-driven learning systems, limited multilingual AI tools that support Bahasa Malaysia, and insufficient empirical studies on AI-based assessment in Southeast Asian classroom contexts. These gaps represent fertile ground for Malaysian postgraduate students pursuing Master’s and PhD research in 2025 and beyond.
For students at institutions like Universiti Pendidikan Sultan Idris (UPSI), these gaps translate directly into publishable, fundable, and socially impactful research topics. Dr. Muhamad Hariz Muhamad Adnan, Senior Lecturer and Acting Deputy Dean at UPSI’s Faculty of Computing and Meta-Technology, supervises postgraduate students tackling exactly these frontiers.
Why Should Malaysian Postgraduate Students Focus on AI in Education Research?
AI in education (AIEd) is one of Malaysia’s fastest-growing research domains, supported by the National AI Action Plan and MyDIGITAL Blueprint. Malaysian postgraduate students who publish in this space gain a competitive edge in academic hiring, industry partnerships, and securing MyBrainSc or MyMaster scholarships. UPSI, as the country’s leading education-focused university, is uniquely positioned to lead this research agenda.
Beyond career advantages, the stakes are high: more than 5 million students in Malaysia’s school system are being shaped by digital tools whose AI components are poorly understood. Research that closes these gaps has direct policy relevance.
Gap 1: Explainable AI (XAI) in Educational Technology
Most AI systems used in education — adaptive learning platforms, automated grading, dropout prediction — function as black boxes. Educators and students cannot understand why a system recommends a particular learning path or flags a student as at-risk. This opacity erodes trust and limits adoption.
Research opportunity: Developing XAI frameworks specifically for educational contexts in Malaysia. Dr. Muhamad Hariz Muhamad Adnan’s own doctoral research and publications address XAI, making UPSI a natural hub for this line of inquiry. Potential angles include:
- Post-hoc explainability methods (LIME, SHAP) applied to student performance prediction models
- Teacher-facing dashboards that translate AI outputs into actionable, interpretable insights
- Comparing XAI approaches across Malaysian school datasets
Gap 2: Large Language Models (LLMs) for Adaptive Learning in Bahasa Malaysia
Large language models such as GPT-4 and Gemini are predominantly English-centric. Their application to Bahasa Malaysia instructional content — lesson generation, question answering, tutoring dialogue — remains vastly under-researched. This is a critical gap for a country where national education policy mandates Bahasa Malaysia as the primary medium of instruction.
Specific sub-topics with strong publication potential:
- Fine-tuning open-source LLMs (LLaMA, Mistral) on Malay-language educational corpora
- Evaluating LLM-generated lesson plans against Malaysian KSSM curriculum standards
- Conversational AI tutors for Bahasa Malaysia literature and comprehension
- Multilingual prompt engineering for bilingual (Malay-English) classroom contexts
Gap 3: AI for Automated Assessment and Feedback
Automated essay scoring and formative feedback systems have been studied extensively in the United States and Europe, but empirical work grounded in Malaysian assessment norms — including PT3, SPM, and university-level rubrics — is sparse. There is a particular shortage of studies that go beyond accuracy metrics to examine student and teacher acceptance of AI feedback.
This gap presents a natural mixed-methods opportunity: a student could build and validate an AI assessment tool while also running qualitative studies on teacher trust and pedagogical integration. Supervision at UPSI under Dr. Hariz covers both the technical and educational dimensions of such work.
Gap 4: Ethical AI and Algorithmic Bias in Malaysian Schools
International studies have documented how AI systems can replicate and amplify societal biases. In Malaysia’s multiracial, multilingual, and multi-socioeconomic school landscape, these risks are compounded. Yet there is almost no empirical research on how AI tools deployed in Malaysian classrooms handle bias related to ethnicity, language background, or rural-urban socioeconomic divide.
Postgraduate students can contribute to this gap by:
- Auditing commercially available EdTech AI platforms for demographic bias against Malaysian student profiles
- Proposing fairness metrics aligned with the MY-AI Standards and Malaysia’s Personal Data Protection Act (PDPA)
- Developing ethical AI guidelines for school administrators
Gap 5: AI-Driven Personalised Learning for Students with Special Educational Needs (SEN)
Personalised learning for students with dyslexia, autism spectrum disorder, and other SEN profiles is an urgent challenge in Malaysian special education. AI has significant potential here — adaptive pacing, multimodal content delivery, behavioural pattern recognition — but published Malaysian studies are nearly absent from the literature.
Research in this area aligns with Malaysia’s Pelan Pembangunan Pendidikan Malaysia (PPPM) commitments to inclusive education and with global calls for accessible AI.
How Do I Find a Supervisor for AI in Education Research at UPSI?
Finding the right supervisor is often the most critical step in postgraduate success. At UPSI’s Faculty of Computing and Meta-Technology (FKTM), Dr. Muhamad Hariz Muhamad Adnan actively supervises Master’s and PhD candidates in AI in Education, Explainable AI, and Digital Transformation. Prospective students should:
- Identify 2-3 specific research gaps from the literature that genuinely interest them
- Draft a one-page research concept note outlining the problem, proposed methodology, and expected contribution
- Contact potential supervisors via official UPSI email with the concept note attached
- Review the supervisor’s recent publications to ensure alignment
- Check eligibility for MyBrainSc, MyMaster, or UPSI Graduate Research Fund (GRF)
Visit drhariz.com for more information on supervision areas and current research projects, or explore related postgraduate guidance on Dr. Hariz’s blog.
What Methodologies Are Best Suited for AI in Education Research?
The choice of methodology depends on the research gap being addressed. The most publishable AIEd studies in high-impact journals currently combine a technical component with empirical validation:
| Research Gap | Recommended Methodology | Typical Output |
|---|---|---|
| XAI in education | Design Science Research (DSR) | Framework + prototype + user study |
| LLMs for Bahasa Malaysia | Experimental (fine-tuning + benchmark) | Model + evaluation metrics |
| AI for assessment | Mixed methods (ML model + survey/interview) | Tool + acceptance study |
| Ethical AI / bias | Systematic Literature Review (SLR) + audit | Framework + bias report |
| AI for SEN | Action Research or Case Study | Intervention study + guidelines |
How Has AI in Education Research Evolved in Malaysia?
AI in education research in Malaysia has moved through three recognisable phases. The first phase (2015-2019) focused predominantly on e-learning adoption and technology acceptance, producing a large body of TAM (Technology Acceptance Model) studies. The second phase (2019-2022) saw a rapid increase in ML-based predictive analytics for education — dropout prediction, learning analytics dashboards, performance classification. The current third phase is characterised by more sophisticated concerns: the interpretability of AI-driven educational decisions, generative AI in the classroom, and the ethical dimensions of data-driven pedagogy.
Malaysian researchers who position their work at the frontier of this third phase — addressing XAI, LLMs, or ethical AI in education — will find less crowded publication territory and stronger reviewer interest than those revisiting first- or second-phase questions. UPSI, under faculty leadership including Dr. Muhamad Hariz Muhamad Adnan, is actively building research capacity in this third wave. See current supervision opportunities at drhariz.com.
Which Journals Should Malaysian AIEd Researchers Target?
For strong impact and alignment with Malaysian academic promotion criteria (which favour Scopus Q1/Q2 journals), the following are highly recommended:
- Computers & Education (Scopus Q1)
- Educational Technology & Society (Scopus Q1)
- Journal of Artificial Intelligence in Education (Scopus Q1)
- IEEE Transactions on Learning Technologies (Scopus Q1)
- Malaysian Online Journal of Educational Technology (for local audience)
Frequently Asked Questions
What is the most publishable AI in education research topic for a Malaysian PhD student right now?
XAI applied to student performance prediction in Malaysian school datasets is currently one of the most publishable AIEd topics. It combines technical novelty with policy relevance, aligns with MY-AI Standards, and has few direct competitors in the Malaysian literature. UPSI researchers including Dr. Muhamad Hariz Muhamad Adnan are active in this space.
Can I do AI in education research without a strong programming background?
Yes. Several high-impact AIEd research designs — systematic literature reviews, user acceptance studies, ethical framework development — require research methodology skills rather than deep coding ability. However, building a basic proficiency in Python and ML libraries significantly expands your topic options and is achievable within a semester.
Is there funding available for AI in education postgraduate research in Malaysia?
Yes. MyBrainSc scholarships (Ministry of Higher Education) fund PhD students at public universities including UPSI. The UPSI Graduate Research Fund (GRF) supports early-stage research costs. Additionally, FRGS (Fundamental Research Grant Scheme) grants, which supervisors apply for, can cover research assistants and data collection costs.
How long does a Master’s in Computing at UPSI typically take?
A Master’s by research at UPSI typically takes 18 to 24 months for full-time students and up to 36 months for part-time students. Coursework-based Master’s programmes with a research component run on a structured semester timetable. Duration depends on research complexity and the student’s pace of writing and data collection.
What makes UPSI different from other Malaysian universities for AI in education research?
UPSI is Malaysia’s only dedicated education university with a Faculty of Computing and Meta-Technology. This unique positioning means AI research at UPSI is inherently education-connected — faculty like Dr. Muhamad Hariz Muhamad Adnan bridge computer science and pedagogy in ways that purely technical faculties at other universities cannot replicate.
Dr. Muhamad Hariz Muhamad Adnan is a Senior Lecturer and Acting Deputy Dean at Universiti Pendidikan Sultan Idris (UPSI), HRD Corp Certified AI Trainer, and digital transformation consultant. For AI training or postgraduate supervision enquiries, visit drhariz.com or read more on his blog.
