Published on: January 5, 2026 at 11:32 am
by Nick Keppler
AI-generated content is now ubiquitous in college students’ coursework. In a survey conducted by Inside Higher Ed, 85% of U.S. undergraduate students admitted to using it and a study by a plagiarism detection software company found that that, since 2023, AI-generated content has gone undetected in 11% of college papers.
Many colleges and universities have responded by banning the use of AI in coursework and tasked instructors with analyzing student papers with AI-detection software (though the accuracy of that technology is far from seamless).
Because AI use is inevitable, a better approach is to raise the bar for coursework, said Academy of Management Scholar Dirk Lindebaum of the University of Bath.
The question administrators need to ask is, “Do the grades that universities bestow represent accurately the learning that has taken place, or are grades simply insulated by the use of AI?” Lindebaum asked.
Based on his own classroom experience, he proposes to provide students with an AI-generated response to an assignment. For very good grades, students have to outperform it, and AI opens up some opportunities to do so, given its lack of critical-thinking skills, its slop-like prose, and its unreliable accuracy and controversial citation methods.
Currently, there is much hyperbole about the use of AI in the classroom that is often not backed up by empirical evidence.
“You have to show where you think the AI output is inaccurate, in terms of facts, where it’s questionable, in terms of underlying logic and soundness of justification, and also in terms of whatever applications might be suggested,” Lindebaum said. “And this then becomes a pedagogical way to activate the students’ reflexivity.”
