Does AI actually help students learn? A recent experiment in a high school provides a cautionary tale.
Researchers at the University of Pennsylvania found that Turkish high school students who had access to ChatGPT while doing practice math problems did worse on a math test compared with students who didn’t have access to ChatGPT. Those with ChatGPT solved 48 percent more of the practice problems correctly, but they ultimately scored 17 percent worse on a test of the topic that the students were learning.
A third group of students had access to a revised version of ChatGPT that functioned more like a tutor. This chatbot was programmed to provide hints without directly divulging the answer. The students who used it did spectacularly better on the practice problems, solving 127 percent more of them correctly compared with students who did their practice work without any high-tech aids. But on a test afterwards, these AI-tutored students did no better. Students who just did their practice problems the old fashioned way — on their own — matched their test scores.
Like any tool, it depends how you use it. I have been learning a lot of math recently and have been chatting with AI to increase my understanding of the concepts. There are times when the textbook shows some steps that I don’t understand why they’re happening and I’ve questioned AI about it. Sometimes it takes a few tries of asking until you figure out the right question to ask to get the right answer you need, but that process of thinking helps you along the way anyways by crystallizing in your brain what exactly it is that you don’t understand.
I have found it to be a very helpful tool in my educational path. However I am learning things because I want to understand them, not because I have to pass a test and that determination in me to want to understand is a big difference. Just getting hints to help you solve the problem might not really help in the long run, but it you’re actually curious about what you’re learning and focus on getting a deeper understanding of why and how something works rather than just getting the right answer, it can be a very useful tool.
Why are you so confident that the things you are learning from AI are correct? Are you just using it to gather other sources to review by hand or are you trying to have conversations with the AI?
We’ve all seen AI get the correct answer but the show your work part is nonsense, or vice versa. How do you verify what AI outputs to you?
You check it’s work. I used it to calculate efficiency in a factory game and went through and made corrections to inconsistencies I spotted. Always check it’s work.
Exactly. It’s a helpful tool but it needs to be used responsibly. Writing it off completely is as bad a take as blindly accepting everything it spits out.
I mean, why are you confident the work in textbooks is correct? Both have been proven unreliable, though I will admit LLMs are much more so.
The way you verify in this instance is actually going through the work yourself after you’ve been shown sources. They are explicitly not saying they take 1+1=3 as law, but instead asking how that was reached and working off that explanation to see if it makes sense and learn more.
Math is likely the best for this too. You have undeniable truths in math, it’s true, or it’s false. There are no (meaningful) opinions on how addition works other than the correct one.