With AI’s ability to solve complex math problems in a matter of seconds, it may feel to teachers like the technology is rapidly changing—or will soon—how math is taught.
When free and widely available tools can solve complex word problems, what math do students need to learn? How can teachers leverage this technology in their math classes while avoiding potential risks?
Education Week put these questions to Jeffrey Bush, an assistant research professor at the University of Colorado Boulder in its Institute of Cognitive Science. Bush studies the use of technology in K-12 STEM education with a focus on math and AI. He’s also a former high school math teacher.
Math teachers may be better prepared to meet this moment than they think. Students still need a strong foundation in core math concepts, Bush said, and this is a subject that has a history of weathering technological disruption.
This interview has been edited for length and clarity.
How should teachers integrate AI into math instruction in ways that are meaningful and age-appropriate?
In terms of lesson prep, I’ve talked with a number of teachers, and I’ve heard some really interesting and creative applications through [my research]. One that comes to mind is an instructor wanted examples of what types of misconceptions or emerging understandings should they expect from this type of problem.
These generative tools can automatically come up with lots of examples and suggestions. You can also turn to them as a coach [and ask], what are some other ways to explain this to someone?
The other side is putting it in students’ hands, which I haven’t seen concrete data on, but from my experience in the math classrooms that I have been in in the last 15 years, even before generative AI, these things [advanced technologies] are not as often put in the hands of youth. I didn’t experience anything more than a graphing calculator until I was in the second year of a mathematics minor in undergrad.
So, there’s lots of ways that AI can be leveraged, and I think it’s really effective when teachers are explicit: “Here’s an AI assignment, here’s what I’m asking you to do on AI. I want you to share what your prompts were, what your outputs were, how you re-engineered your prompt.”
It’s not a nefarious back shortcut. It’s a learning tool that’s being brought into the classroom, and we’re learning how to use responsibly. That promotes AI literacy, and it also creates a transparent classroom environment.
What are the potential pitfalls of using AI in math instruction?
Inaccuracy, particularly inaccuracy that comes with bias. Both have the potential to do harm. And in an era where information is becoming harder to verify than it has been in the past, AI seems to provide expertise and rigor but hides some of the mistakes and biases behind its reputation. I’ve had plenty of instances myself in doing research—if I look to AI and then I verify things, I can sometimes be quite surprised at the confidence with which information is presented that is wrong.
In addition to that being a concern, I think it’s a call to action around AI literacy. We need to be able to understand, how do we check the sources? How do we know where these things come from? What are the types of things that we should and shouldn’t be asking AI, or should be asking and skeptical about? That’s extremely important and can help mitigate some of the problems.
Are you concerned that with AI’s abilities, kids will come to see math as something that is done by robots, not humans?
I think the position that math is in—that’s unique compared to social studies, English/language arts, and computer science—is that robots have been doing math for a long time, and there’s a deep and long history of the automation of mathematical products and progress over time. And these AI tools have new capacities, especially capacities around writing and programming, but also new capacities around math.
I think is a great touch point in this, an example in math history, is the film “Hidden Figures.” The characters in that movie are asked to be computers. They were doing calculus. Now, all of that work is automated, but at the time, in the era of the space race, people doing that work was extremely important. Being able to do the math and understand what’s going on sets those people up for upskilling and reskilling. It’s not like once a computer was invented, they were out of work. They had great skills. And knowing how to do calculus means you can understand change, and you can understand complicated systems, and you have confidence around solving problems that applies to a lot of other contexts.
I think be transparent about why we’re doing what we’re doing. [Students] know when they’re being deceived and told, “you’re going to have to do this long division when you grow up.” They’re probably not doing long division, but understanding the base 10 number system is an important way to develop your brain to learn how to think.
Understanding the foundations—the number sets, the mathematical literacy—is more important than ever because this critical thinking is something that AI has not yet been able to replace in humans.
