Presented by
Session Description
A 2024 controlled study at a U.K. university submitted fully AI-generated papers under fake student profiles. Faculty failed to identify 97 percent of them (Turnitin, 2024). The Digital Education Council's 2025 global survey found that 88 percent of students have used generative AI for coursework, up from 53 percent the prior year. And yet only 39 percent of institutions have an AI-related acceptable use policy in place (EDUCAUSE, 2025).
These numbers suggest something more uncomfortable than a cheating problem. They suggest that many of our most trusted assessment practices were measuring the wrong thing all along. We built systems around artifacts that were easy to collect and easy to defend to accreditors, and for decades that worked well enough. AI has not disrupted the system so much as it has revealed what the system was actually rewarding; namely, polished output, efficiently produced, with no reliable connection to the thinking we claimed it represented.
