Generative AI as a Catalyst for Self-Regulated Learning: Harnessing Multimodal Data to Transform Education
Roger Azevedo. University of Central Florida (USA)
EDULEARN26
Artificial intelligence (GenAI) is opening new possibilities for understanding how students learn, think and regulate their own learning processes. In this talk, Professor Roger Azevedo explores how GenAI can support self-regulated learning in technology-rich educational environments. Drawing on contemporary theories of metacognition and learning sciences, he examines how multimodal data — including learners' interactions, discourse, gaze and physiological signals — can help model students' cognitive, affective and motivational processes in real time. The talk highlights the potential of GenAI-powered pedagogical agents to provide adaptive, proactive and explainable support tailored to individual learners' needs. At the same time, it addresses key challenges related to validity, transparency, bias and the ethical use of learner data. Professor Azevedo outlines a forward-looking agenda for building intelligent systems that not only assess learning, but also help students plan, monitor and reflect on their own learning across diverse educational contexts.