Using generative AI to build discovery interview capacity: Lessons from a Canadian college

Abstract

Training staff to conduct discovery interviews for potential academic misconduct can be challenging, particularly for institutions with limited numbers of experienced interviewers, geographically dispersed campuses, or growing demand for academic integrity support. To address this challenge, the author developed and piloted a generative artificial intelligence (genAI)-supported discovery interview simulator designed to provide staff with opportunities to practise interview techniques, receive immediate feedback including alignment with policy and procedure, and build confidence in a safe learning environment. Rather than replacing traditional professional development, the simulator was intended to complement existing training by expanding opportunities for realistic, policy-aligned practice. 

Drawing on participant feedback and implementation experience, this article describes the development of the simulator, lessons learned during its pilot use, and considerations for institutions interested in exploring similar approaches. Participants reported positive experiences with the simulations, particularly valuing the opportunity to practise realistic conversations, receive formative feedback, and engage in low-risk experiential learning. The pilot also highlighted the importance of aligning genAI-supported training with institutional policies, refining scenarios through ongoing iteration, and integrating genAI-based simulations with facilitated discussion and mentorship. 

This experience suggests that genAI can serve as a practical tool for expanding institutional capacity for discovery interview training while maintaining human oversight and educational purpose. Recommendations are offered to help Canadian, and other, postsecondary institutions evaluate whether genAI-supported simulations may be appropriate within their own academic integrity training programs.