Center for Innovation in Teaching & Learning

In an extreme challenge to our educational mission, instructors are concerned — and in some anecdotal cases, already convinced — that students are simply using generative AI to produce all their course deliverables, and that no learning is actually taking place. In this talk, we will report on our effort to test this hypothesis: is it possible to "excel" in a course without absolutely any knowledge? We show that, while there is variability across courses and assignment types, large language models are indeed often able to achieve a strong grade in a course without any meaningful help from human users or course material. This finding naturally brings up questions about our future assessment methods, and even about the value of the material that we are teaching: while obviously yet unresolved, we will describe some possible ways forward at the end of the talk.