Reposted by Brittany R. Clark, Ph.D.
Sobering but also reinforces that solving any of these problems is going to be structural. As long as emphasis is on grades and doing as much as possible, as long as universities call for "embracing" AI, as long as assessment trumps learning, it's going to be hard to solve this even as many/1
![He understands why this shift has happened. AI use is “very terrible” at Penn, he said. It’s not just blatant cheating that’s a problem. Bringing assignments back into class helps, but students “are still going to use AI to plan out what they’re going to write, and they’re still going to use AI to summarize the readings they’re supposed to do at home,” he said. “I’m not saying every student is like this, but [in-class work] is a Band-Aid basically to a much larger problem, because critical thinking is still largely being outsourced.”
Pedersen said he has been as guilty as his classmates of overrelying on AI. He recalls not paying attention in an introductory course, asking AI to create summaries of lectures and notes, then bombing a test. Last year, in a course on European politics and institutions that he took while studying abroad, students were given 10 essay prompts in advance of the in-class final exam and told they would have to respond to three. Feeling overwhelmed by the number of readings he had to keep track of and running out of time, he turned to AI to do much of the preparation. He received an A.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:a6m5puxjmwuwxhuviabkhg67/bafkreiahwluei27sitsb3u4faeihmjj3xr64ezwhhblb6uus2fx4o6tpum)
