A game-theory professor ran a take-home exam on the honor system. 50 students used AI

A 34-year career's first take-home exam met its first mass-cheating scandal

The El País report profiles a Brown economist — a game-theory researcher — who for the first time in 34 years set both his midterm and final as take-home, closed-book exams, leaning on the Ivy League tradition of trusting students. His reasoning was pedagogical: "It's a very nice kind of exam, because as you're giving students practically unlimited time to complete it, it lets you make it harder than normal, to see how far they can go." The course drew 86 students, far above its usual ~30.

The trap sprang itself. The final was in person, and the score distributions between the take-home midterm and the proctored final diverged so sharply that he discarded the midterm grades entirely. He says he has conclusive evidence at least 50 students cheated on the March midterm, calling it the biggest known scandal at Brown and across the Ivy League. The bitter irony he names himself: "We economists understand reality as a set of people responding to optimization problems with restrictions" — and he warns the stakes run to "the future of higher education." When he tried to escalate, his own department offered no response.

"A game theory expert can't see the problem here"

Hacker News did not extend much sympathy. The most-repeated objection was that the professor engineered his own disaster, and that the people best positioned to know it were the ones piling on:

"His research is in Game Theory. He should have realized that, in a situation where all competitors are (possibly) using LLMs, the game theoretic optimal choice is to use LLMs." — a commenter on the game-theory irony

"take-home, closed-book type — What an oxymoron." — userbinator, 90k karma

A different voice flipped the blame back onto the system the students inhabit, arguing the honor code is a luxury good that died with cheap tuition:

"Back in your day education was not sold as a financial transaction to a prosperous life. The honor system… requires idyllic circumstances: people who pursue education for no reason other than curiosity… It's not 'kids these days', it's the natural result of the systems adults have made." — a commenter rebutting an appeal to the honor code

A Dartmouth instructor is redesigning the course as warfare

The most concrete reversal came from inside the same fight. A Dartmouth CS instructor described rebuilding their intro systems course from scratch — not to ban AI, but to make grade-optimization and learning the same path:

"we're thinking of it as an adversarial problem… we're designing the course to ensure that a student optimizing for the best grade per unit work still meets our learning objectives." — a CS instructor at Dartmouth

That means paper exams plus one-on-one interviews where a student must explain a macro they used or why their code is structured as it is. His framing was pointed: "It's best not to blame the students. They are good at optimizing metrics; that's how they ended up here." A retired science professor turned high-school AP teacher confirmed the scale from the trenches, calling the cheating problem "a crisis greater than COVID," with most students in advanced classes cheating and many teachers having simply given up.

What the thread surfaced that the article didn't: the proctored final wasn't just a detection method, it was the only assessment in the course that survived contact with reality. The professor's mistake wasn't trusting students — it was building a measurement the smartest move was to corrupt, then acting surprised that economically rational people corrupted it. The fix every working instructor in the thread converged on costs the one thing universities priced out long ago: a human in the room, one student at a time.