Project Team
Robert Hill, Physics and Astronomy
Jason Thompson, Integrated Teaching Support Unit
Stephanie Boragina, Integrated Teaching Support Unit
Kristin Wilson, Integrated Teaching Support Unit
Yasin Dahi, Integrated Teaching Support Unit
Vivian Dayeh, Biology
Steve Forsey, Chemistry
Niayesh Afshordi, Physics and Astronomy
Megan McCarthy, Psychology
Darren Gigliozzi, Optometry
Project timeline: September 2026–February 2028
Description
This project will provide students in six STEM courses at UWaterloo access to Axiom, a GenAI platform that houses AI learning agents. Each agent is tailored to a specific instructional activity and constrained by prompts that provide pedagogical guardrails and prioritise course materials. The goal is to leverage the affordances of GenAI, while ensuring appropriate cognitive engagement that will support student learning.
GenAI use is growing, as a recent survey, shows 65% of UW students reported using AI tools for academics (co-applicant1, 2025). Research shows that indiscriminate use of GenAI can harm student learning, while constrained use may be supportive. Informed by the cognitive science of learning, we propose that the impact of GenAI depends on whether intrinsic learning processes are bypassed or reinforced—i.e., how students use it in their learning.
We will assess the efficacy of our GenAI intervention by measuring student performance and perceptions, and instructor experiences.
Project Goals
Our principal goals are to investigate whether GenAI can be used to enhance student learning by,
- providing effective, individualised feedback on student work created as part of in-class learning activities (i.e., feedback agent).
- helping students to prepare for summative assessments (i.e., self-testing agent).
- supporting students and instructors by providing accurate and timely information on course administration (i.e., admin agent).