AI enters Princeton, where 30% of students cheat – but their peers don’t.

Even after July 1, however, professors cannot directly interfere with attempted plagiarism. Instead, they will observe and take notes, serving as an “extra witness in the room” who can testify in cases that will later be brought to the Honorable Court.

AI has rapidly improved education, forcing many teachers to abandon the task of writing and take tests at home with in-class or oral exams. As the Princeton example shows, even this is not enough; most students, given the opportunity, will happily use AI in the classroom surrounded by their peers if they can get away with it.

Such a spread of ideas and memories is very frustrating for many teachers. This includes our own Scott Johnson, who he recently wrote a piece for Ars about what it feels like to post more machine-made answers rather than human-made ones. (Note: Nothing good.)

It’s not like students think they’re learning while doing this; they are too smart for it, especially at Princeton. But while the pressure to succeed remains high, and the cost/complexity of AI tools remains low, many students are tempted to take a shortcut, even one that hurts them. As Scott concluded:

I have never met students who think they are learning when they let LLMs do their work, even though college administrators and LLM trades try to impose this. It’s job management for them.

Who knows what will happen if the AI ​​bubble bursts and the easy and ubiquitous access to the LLM dwindles into something much smaller. But while AI is here, it is not changing education and improving education. It just makes it more difficult to do all the things that have helped students learn for so long.

That’s not how the AI ​​industry sees things, of course. After reading this week’s Daily Princetonian article on proctoring reform, I couldn’t help but notice the big flag at the top of the piece. “HE TRIED TO GET READY,” it said. It was a Google Gemini product.