Remember when everyone was amazed by ChatGPT? That's so 2022.
Political scientist David Schultz just concluded his semester-long experiment at Hamline University in St. Paul that allowed students to use ChatGPT and other AI chatbots for their coursework.
The result?
"No kid is wowed by it," Schultz said. "That's what is so interesting to me. Nobody is like, 'I gotta use this stuff all the time.'"
College-age adults have already seen a lot of fascinating innovations in their young lives. Cars that are nearly autonomous, rockets that take off and land on the same launch pad and any piece of information or entertainment at their fingertips. It's all normal to them.
And so AI chatbots — and the language modeling they're based on — amount to just another new tool to them. None of the students I watched in two of Schultz's courses this semester found that ChatGPT could substitute for their own thinking and writing.
"It's a really strong language model and that's really impressive," Ian Creech, a student in Schultz's Ethical Policy class, said. "It's not generating original thought. It just pulls it from everywhere else."
In many ways, I'm heartened by the modest impact AI had on the Hamline students. They weren't able to cheat their way through school with it. And it didn't impose conclusive ideas to the knotty tradeoffs Schultz taught them, a limitation that appears to be deliberate in ChatGPT but that may be different in future types of AI systems.