Guest editorial: Harnessing intelligent technologies for personalised learning experiences

The paper “From Emotional Safety to Agentic Engagement: Implementing Mentimeter in Undergraduate TESOL Courses” by Yan et al. addresses a persistent challenge in English-medium instruction contexts for non-English-speaking learners: student passivity and language anxiety. Using Mentimeter as a student response system, the authors investigate how anonymity, low-stakes participation and structured interaction can help learners move from emotional safety towards greater agency and cognitive engagement.

A notable contribution is the proposed staged engagement framework. Rather than treating engagement as a monolithic construct, the authors argue for a sequenced design: beginning with anonymous, low-pressure activities such as quiz games to establish psychological safety, then gradually scaffolding towards more cognitively demanding tasks such as open-ended critique and reflection. This framework offers practical guidance for educators seeking to harness student response systems to cultivate critical, agentic participation in technology-mediated classrooms.

Personalisation at the level of content and conceptual understanding is the focus of “Personalised Smart Courses for University Physics Based on Knowledge Graph and Large Language Models” by Zhang et al. The authors present the Physics Intelligent Teaching Assistant, an LLM-based platform that integrates retrieval-augmented generation and graph neural networks for knowledge tracking, aimed at advancing conceptual mastery in college-level physics.

This intelligent assistant is designed to support adaptive learning, student-tailored dialogue and personalised resource recommendations by tailoring explanations, decomposing complex problems and adapting dialogue to student learning styles. It enables tracking of conceptual mastery to dynamically adjust instructional strategies. The work demonstrates how combining domain-specific knowledge bases, adaptive dialogue and interpretable knowledge tracking can foster more robust problem-solving skills and conceptual understanding and promote critical thinking and independent inquiry.

The role of immersive technologies in supporting performance-based skills is explored in “Virtual Reality for Presentation Skill Mastery: Measuring Knowledge, Performance and Confidence with Chill Talk” by Chan. Focusing on English as a Second/Foreign Language learners, this paper presents Chill Talk, a virtual reality (VR)-enhanced platform designed to develop public speaking skills through self-directed learning.

The paper adopts a design-orientated perspective: VR is leveraged as part of a constructivist learning ecology that combines accessibility, immersion and flexible practice opportunities. By allowing learners to practise at their own pace, in varied scenarios and with reduced social pressure, Chill Talk illustrates how immersive environments can support personalisation in skill development.

While personalisation often focuses on learner-facing experiences, it also depends on back-end infrastructure capable of matching diverse learners with appropriate resources. “A Proposed AI/ML Algorithm for Optimising the Selection of Educational Resources” by Al Ka’bi et al. addresses this challenge by proposing an algorithm tailored to process and classify heterogeneous digital learning resources to support accurate personalisation. The algorithm tackles two common problems in educational recommender systems: data sparsity and information loss in long-form, multimodal content.

Empirical results suggest improvements in recommendation precision and convergence stability while maintaining response times suitable for interactive systems. This recommendation algorithm shows potential to contribute to more equitable access to appropriate learning materials, especially in environments where content overload and limited guidance pose significant barriers to effective learning.

As generative AI tools rapidly diffuse into higher education, questions about their impact on higher-order thinking skills have become critical. “Explore the Gen-AI Empowerment Black Box: A Meta-Analysis of the Impact of Gen-AI on College students’ Critical Thinking” by Jiang offers a timely and systematic synthesis of empirical studies examining this relationship. By delineating critical thinking from broader notions of higher-order thinking, the meta-analysis provides a focused assessment of how Gen-AI interventions influence critical thinking outcomes.

The study identifies five moderating variables of the use of GenAI on critical thinking: discipline, knowledge type, pedagogical approach, GenAI role and task type. GenAI has the greatest positive impact on critical thinking in STEM disciplines, for procedural knowledge, within inquiry-based learning designs, when positioned as a peer and for tasks emphasising reflection and metacognition. These findings highlight that the benefits of GenAI are contingent on how tools are framed, integrated and scaffolded within pedagogical designs.

Personalisation encompasses not only cognitive but also affective aspects of learning. “A Neuro-Symbolic Reasoning for Affective-Aware Personalisation in Virtual Learning Environments” by Olaniyan and Wario introduces an affect-aware learning agent that combines deep learning with symbolic reasoning to adapt instruction based on learners’ emotional states. The system integrates multimodal emotion recognition with a reasoning engine that adjusts instructional strategies in real time.

A key contribution lies in integrating transparent, interpretable rules with neural models. This hybrid design supports emotionally responsive and pedagogically meaningful interventions while maintaining a degree of explainability. The proposed system highlights the potential neuro-symbolic approaches to enhance learner engagement, promote self-regulated learning and support personalised instruction in virtual learning environments.

Personalisation also entails improving the broader teaching and learning ecosystem through better use of data. In “Multi-dimensional Academic Analytics for Course Evaluation”, Wong and Li present an institution-wide analytics platform that integrates student feedback, course performance data and contextual information to support more nuanced course evaluation and quality assurance.

Moving beyond traditional survey-centric approaches, the platform offers descriptive, trend and pattern analyses at the levels of course, instructor and academic term, presented through user-friendly dashboards. By enabling users to relate satisfaction scores to performance patterns and contextual factors, the platform supports reflective teaching and data-informed decision-making without requiring advanced data literacy. This work underscores that personalisation is not limited to individual learners; it also involves tailoring institutional interventions, support structures and professional development to the specific needs evidenced by data.

These papers collectively present a rich and multifaceted view of how intelligent technologies can enable personalised learning experiences across contexts, modalities, disciplines and scales. The studies in this special issue suggest several promising directions for research and practice:

• Integrating multimodal, affective and behavioural data streams in ways that respect privacy and learner agency, while supporting richer forms of personalisation.

• Exploring long-term impacts of AI-enabled personalisation on learner autonomy, critical thinking and lifelong learning habits.

• Developing design frameworks that help educators use AI tools, immersive environments and analytics platforms coherently, rather than as isolated interventions.

• Advancing methods for evaluating AI-driven personalisation, not only in terms of performance gains but also in relation to ethical, motivational and socio-cultural outcomes.

By showcasing diverse yet complementary approaches to intelligent, personalised learning, this special issue aims to support educators, researchers, policymakers and technology developers in navigating a rapidly evolving landscape. We hope that the insights, frameworks, platforms and empirical evidence presented here will inspire further innovation and critical reflection and contribute to more engaging, inclusive and responsive educational experiences for learners.