Guest talk: Michael Liut "The Wonderful World of AI: Adaptive Systems, Adaptive Content, Adaptive Feedback for a Unique Student Experience"

The Wonderful World of AI: Adaptive Systems, Adaptive Content, Adaptive Feedback for a Unique Student Experience

Michael Liut
University of Toronto Mississauga (UTM)

Abstract: Computer science classrooms are stubbornly heterogeneous: students arrive with wildly different prior knowledge, motivational dispositions, and ways of engaging with material, yet most receive the same lectures, the same problems, the same feedback, and the same visualizations. This talk will focus on generative AI — spanning both Small and Large Language Models — finally making it tractable to move past one-size-fits-all instruction and toward genuinely adaptive learning experiences designed around the individual student.

I will focus on three intertwined threads of this work: (i) Adaptive Systems that model learners along cognitive and motivational dimensions and orchestrate what each student sees next; (ii) Adaptive Content that personalized problems, worked examples, and even custom visualizations that calibrate scaffolding, explicitness, and representation to the learner rather than to a hypothetical median student; and (iii) Adaptive Feedback where we can detect common misconceptions as they emerge, providing learner-appropriate explanations, and enabling custom SLM/LLM-driven grading that returns substantive, personalized commentary at a scale no instructor could reach alone. 

Finally, I will close by sketching what comes next: tighter integration of Small and Large Language Models so feedback arrives in the moment of struggle (not after it!), the need for improved learner personalization, and the open question of whether keeping students engaged actually translates into lasting understanding. 

The Wonderful World of AI in education, is certainly not about replacing instructors or students, rather it is about finally giving every student a learning experience shaped around them. :)

Bio: Dr. Michael Liut is an Assistant Professor, Teaching Stream in the Department of Mathematical and Computational Sciences at the University of Toronto Mississauga (UTM), where his work sits at the intersection of computer science and education. His research centres on the design of adaptive systems that leverage personalization through both small and large language models to enhance the educational experience; drawing on a background in learning sciences and engineering, algorithmic design, and optimization. Building on earlier work developing predictive and prescriptive models in higher education, Michael now focuses on the creation of AI-driven tools, behavioural interventions, and adaptive experimentation frameworks — such as QuickTA, MedBot, NUMI, and VoiceEx(plantations) — that have supported thousands of students at key moments in their learning. His sustained contributions to student learning and educational innovation have been recognized with the UTM Dean's Merit Award four times in the past five years, UofT's Outstanding Faculty Guidance & Support Award (2021), and most recently, UofT's Early Career Teaching Award (2025-2026).

This guest talk is hosted by Assistant Professor Juho Leinonen, Department of Computer Science.