AI Cheating Spreads to In-Person Classes in the US

New research in the US suggests that digital tools and artificial intelligence are making it easier for students to cheat not only in online courses but also in traditional in-person classes. An analysis by biology professor Sara Brownell at Arizona State University found that, on average, 45% of course points in 21 biology classes could potentially be earned through digital cheating methods.

Brownell said recent education trends have shifted away from relying solely on high-stakes exams and toward rewarding class participation and active learning. However, the widespread availability of tools such as ChatGPT means students can quickly generate answers to assignments or participation questions, raising concerns about whether grades still reflect genuine learning.

The study also found students using a variety of tactics, including responding to in-class questions remotely through electronic clickers, sharing answers in group chats and discreetly photographing test materials to look up answers online. Brownell argues universities may need to rethink assessment methods, including verifying participation more carefully or returning to in-person, pen-and-paper exams in some cases.