On bot-proofing
Michael Cournoyea, Sarah Seeley · Assessment & Evaluation in Higher Education · 2026
As generative AI becomes increasingly adept at solving problems, composing essays, and completing assignments, educators are confronting new pedagogical and social complexities. Many are reimagining assessment strategies to foreground distinctly human capacities while moving away from tasks that can be easily completed by or with such tools. We introduce the concept of bot-proofing to describe deliberate pedagogical approaches to evaluative design that prevent, dissuade, or redirect students from using generative AI for assessments. Drawing on recent research conducted at the University of Toronto, we present a framework detailing four distinct approaches to bot-proofing—avoidant, adaptive, deterrent, and rhetorical—designed to help educators name, describe, and reimagine their pedagogical choices. Bot-proofing strategies prioritise pre-bot learning objectives, which often assume that students must learn in established ways. We conclude by suggesting that the adoption of a strong bot-proofing stance cannot create space to reimagine foundational skills—an exigence that is both pressing and complex.