What were the findings of that two-year study?

So this was a study that looked at a program called Khan Academy, which some listeners may have heard of.

It's designed, it often, it's known especially for its videos, but the program in schools is designed to help students, walk students through math problems and give them math practice for students who are struggling.

And so researchers looked at schools that introduced during the school day Khan Academy programs.

versus schools that didn't or grade levels that didn't introduce Khan Academy and were using other digital products or a small group teacher-led instruction.

The other thing about Khan Academy is it recently introduced an AI chatbot known as Khanmigo, where students, if they were struggling, if they couldn't get the math problem right and the explanations in Khan Academy were not helping them, they could ask the chatbot.

Unlike other AIs, it wouldn't give them the answer.

So what is the best role for AI in American schools? What's the positive? We've talked about some of the potential negatives, but what is the positive?

How did that program benefit students, and what did teachers say about it?

So the program had actually a variety of different aspects of it, including merit pay, but also career advancement opportunities and additional professional development.

And the researchers tracked students in schools that participated in the program, and they found long-term benefits.

Their test scores rose, and they were more likely to graduate high school.

We don't know exactly what teachers thought about the program, but we do know that the sort of general culture in the schools improved.

And so it may have been the case that actually teachers were okay with this particular program, though we don't really know for sure.

Matt, are these big differences in student achievement in these studies? I know some past studies have showed improvement, but in the single digits, and there's a well-known study out of New York City that showed students actually did worse.

worse.In South Carolina, I would describe the difference as modest.

Tell us about that.

So, it depends on who you ask, because for students, they see it as a benefit to get away with not learning, to use it to cheat, essentially. I talked to one 17-year-old high school student in New Jersey who described how she used AI to do a lot of her schoolwork last year. She used it to complete a math take-home quiz. She used it on a science lab. She used it to write a tricky section of an essay. Of course, for harried students who are trying to pass all their classes, it's a great innovation, but it's an innovation in cheating.And for teachers and researchers and even some students themselves, they're worried that students are not learning.

One of the biggest concerns seems to be that this isn't just like us using a calculator to solve a big math problem, for example. This is really students bypassing the entire learning process to learn things like critical thinking skills that they're gonna need later in life. Is there a solution to that, or (laughs) is that something that the companies and educators and parents are really just trying to get to grips with?

That speaks to the larger debate. There are people who say, "No, this is just like the calculator, and we need to find ways to adapt our education system so students are able to use it and take advantage of it." And maybe the learning that they traditionally have done without AI is not as relevant as it used to be, and maybe it's okay if students use AI to generate essays. Now, there are a lot of educators and some researchers I talked to who are very skeptical of that and who say, "You know what? Even if AI is great at generating essays, the process of writing is very helpful for student learning." There is this bigger question of, like, what AI means for education, what it means for the job market, and how, if at all, schools need to adapt.