Fail Students Who Use ChatGPT—or Rethink Teaching? Professors Clash Over AI in the Classroom

As generative AI reshapes education, instructors are divided between strict enforcement and experimentation. Large language models can produce essays, solve problems, and offer ready-made answers, enabling some students to bypass learning—and prompting a range of responses from professors determined to preserve academic rigor.

Two Very Different Approaches

Neal Hebert, a theatre professor at Grambling State University, has adopted an uncompromising stance: he tells students that ChatGPT is prohibited in their writing process and that he can usually tell when a student has relied on AI. "I will fail the student on this assignment if it is used—and, potentially, for the entire course, if we go through a formal appeals process," he told The New Yorker.

"I tell my theatre majors, 'I get paid the same whether I pass you or fail you,'" Hebert said. "What you’ve done is tell us you are so lazy you would rather outsource your collaboration to an app than risk being an artist."

Hebert says the surge of AI-assisted submissions in introductions classes forced him to shift from collaborator to plagiarism cop. After seeing many papers about August Wilson's Fences that shared similar phrasing and a formulaic tone, he now designs assignments around obscure plays that are unlikely to be in AI training data. When AI lacks reliable source material, Hebert says, it often fabricates characters or plotlines, producing unusable results.

An Experimental Alternative

Daniel Silver, a sociology professor at the University of Toronto Scarborough, has taken a different route: he treats AI as a prompt to rethink pedagogy. Silver redesigned assignments to harness AI creatively—asking students to build or test AI agents that simulate historical thinkers—and used conversations to correct misuse.

"AI has fundamentally changed how I teach, and it demands basic reflection about what we are trying to accomplish," Silver told The New Yorker. He reports giving zeros for thoughtless AI use but typically allowing resubmissions after a meeting, which often leads to improvement.

Silver also shows students AI-generated examples to expose formulaic patterns and encourages reflective use of tools rather than uncritical reliance on them. He describes the transition as emotionally taxing but believes it may ultimately strengthen teaching and learning.

The Broader Stakes

Both approaches highlight a central tension: preserving academic integrity versus adapting instruction for a new technological landscape. Hebert worries about the long-term cultural consequences if performers and students stop engaging deeply with texts; Silver sees an opportunity to update learning objectives and assessment methods.

Whatever the method, educators face shared choices: redesign assessments to require original, evidence-based thinking; confront and remediate misuse; or enforce strict penalties to protect standards. The debate continues as schools experiment with policies that balance integrity, creativity, and practical pedagogy in an AI era.

Further coverage examines related controversies, such as parents' concerns about using student video to train AI systems.