Tentative Title:

"A Multimodal AI System for Real-World Collaboration and Regulation of Higher Education Students: Lessons Learned"

Abstract:

Building on the keynote talk, this symposium component will highlight the challenges of AI systems that directly intervene in the teaching and learning processes of social and emotional competencies. To illustrate these challenges, the talk will showcase an AI system designed to support students in socially shared regulation of learning within small group collaborations.

The discussion will explore AI models as tools for examining human learning processes but will take a critical stance on the ability of these models to address complex social and emotional competencies.

In learning analytics, the shift "from clicks to constructs" aims to link digital traces to meaningful educational concepts. The talk will exemplify this by analyzing socially shared regulation of learning through audio and video data, processed via computer vision and NLP, to model group interaction behaviors.

Key points include:

• The risks of over-reliance on mechanistic measures (e.g., timing interactions, response speeds) that might misrepresent learning as a matter of efficiency rather than quality.

• The limitations of AI in fully capturing the complexity of human learning, as these measures are more suited for evaluating machine capabilities than human cognitive and emotional processes.

• A descriptive approach to AI-driven insights into learning, rather than prescriptive automation.

The talk will conclude by exploring how such approaches can foster opportunities for researchers, teachers, and learners in higher education settings.At UCL, he serves as Professor of Learning and Artificial Intelligence and is affiliated with the UCL Knowledge Lab. His work focuses on AI-driven learning solutions, emphasizing the development of skills such as learning how to learn and collaborative problem-solving.

Beyond research, Professor Cukurova actively contributes to the academic community. He is an editor for the British Journal of Educational Technology and an associate editor for the International Journal of Child-Computer Interaction. He also collaborates with UNESCO’s Unit for Technology and AI in Education, shaping policy documents, including UNESCO’s recent report on generative AI in education and research.

His leadership and contributions continue to shape the future of AI in education. For more insights, his keynote address, "Beyond the Hype to Real-world Adoption," offers valuable perspectives

Learning to learn: supporting the affective facet of learning to learn

Many AI systems have supported the affective facet of learning and some have also attempted to support the affective facet of self-regulated learning. 

This talk will describe three kinds of AI system.

1. Purely Reactive systems try to recognise the learner’s affective state and react to it in a way that is hoped to be helpful: for example, by reinforcing a presumed, positive state through praise (say) or by empathising with a presumed, negative state. Such systems will have a limited effect on the learner’s affective self-regulatory capability.  

2. Explicit systems both react to and name the inferred affective state of the learner.   Identifying the affective state to the learner and suggesting how to deal with it during learning should develop the learners’ accuracy in identifying and managing such states. 

3. Pro-active systems specifically engage learners in affectively focused activities during learning to help them learn to manage their affective states both negative and positive.

This talk will discuss all three of the above kinds of system with examples.

Dr. Wannapon Suraworachet is a postdoctoral researcher at UCL Knowledge Lab, University College London, UK, and a member of the UCL Learning Analytics and Technology Team (UCLAIT) led by Prof. Mutlu Cukurova. Her background in computer science and prior experience as an EdTech entrepreneur in Thailand provide a strong foundation for her interdisciplinary research at the intersection of technology and education.

Tentative Title:
AI-Driven Learning Analytics for Equitable and Reflective Collaborative Learning

Abstract:
This symposium presentation will examine how AI and learning analytics can be harnessed to support equitable, reflective, and effective collaborative learning. Dr. Suraworachet’s research is rooted in her motivation to address disparities in access to educational technology—an issue she encountered firsthand in Thailand. Shifting from entrepreneurship to academia, she focuses on designing theoretically grounded, AI-driven tools that enhance learning across diverse contexts.

Her doctoral work at UCL explored how postgraduate students regulate learning in small-group settings, particularly in response to challenge moments. She developed machine learning and language models to detect such moments from student discourse and used them to generate dashboards for reflection. These dashboards offer students evidence-based feedback to support self-, co-, and socially shared regulation of learning. Her contributions are both methodological—advancing how AI can interpret complex constructs like challenge and regulation—and practical, by enabling data-informed educational interventions.

Currently, her postdoctoral work builds on these foundations by incorporating teachers’ perspectives in orchestrating multi-group collaborative classrooms. The goal is to design AI systems that align with teachers’ needs, enhancing teacher-AI collaboration in the classroom rather than replacing teacher agency. This research speaks directly to the evolving expectations of educators in digitally disrupted learning environments.

Dr. Suraworachet actively contributes to the academic community, serving on program committees for international conferences such as the European Conference on Technology-Enhanced Learning (ECTEL) and Learning @ Scale (L@S), and reviewing for respected journals including the British Journal of Educational Technology (BJET) and the Journal of Computer Assisted Learning (JCAL).