Universities rewrite the syllabus after ChatGPT catches on

Last month, Northern Michigan University philosophy professor Antony Aumann was grading essays when he stumbled upon the best paper in the class. It tackled the morality of burqa bans with crisp paragraphs, sharp arguments, and zero fluff. There was just one catch: when confronted, the student admitted that ChatGPT wrote the whole thing.

Now, universities across the U.S. are overhauling the classroom to deal with the sudden arrival of generative AI, transforming how professors teach and assess learning in real time.

Ditching the five-page take-home

While public school districts in New York City and Seattle rushed to ban ChatGPT on school networks, higher education institutions are taking a very different route. University leaders argue that bans simply do not work and risk infringing on academic freedom.

Instead, professors are rewriting the rules of assessment. The take-home, open-book paper—a staple of remote learning during the pandemic—is rapidly being phased out at schools like George Washington University, Rutgers, and Appalachian State.

In its place, faculty are returning to in-class assignments, handwritten blue-book exams, oral presentations, and group projects. Professor Aumann plans to require students to write first drafts inside the classroom using locked-down browsers that monitor activity. For subsequent drafts, students will have to justify every single edit they make.

Outsmarting the chatbot

Assignments are also getting much more personal and esoteric. Generic prompts like "write five pages about this theme" are effectively dead. Professors are pivoting toward questions about students' personal lives and hyper-recent events.

At the University of Texas at Austin, humanities chair Frederick Luis Aldama plans to assign niche texts where AI models have less training data, such as Shakespeare's early sonnets rather than well-trodden plays like A Midsummer Night's Dream. As University of Florida English chair Sid Dobrin bluntly put it, students are plagiarizing because the assignments can be plagiarized.

Meanwhile, OpenAI said it is developing tools to help spot AI-generated text, but with GPT-4 on the horizon and tech giants pouring billions into rival models, schools know this is just the beginning.

Why it matters

We think this classroom panic exposes an uncomfortable truth: university humanities assessments have leaned on formulaic paper-writing for far too long. If an AI can generate an A-grade essay in seconds, it suggests many traditional prompts were testing basic synthesis rather than genuine critical thinking.

Choosing adaptation over outright prohibition is clearly the right call for universities. With Google pushing LaMDA and Microsoft backing OpenAI, trying to ban these tools on campus Wi-Fi is a fool's errand. Forcing students to critically evaluate AI outputs rather than pretending chatbots do not exist is the only sustainable path forward.

Still, retreating into handwritten papers and oral exams looks like a stopgap defense rather than a long-term strategy. It significantly raises grading friction for instructors without solving the deeper question: what skills should students actually be mastering when machines can already draft coherent prose on demand?