As artificial intelligence spreads to a growing number of career fields, experts say it’s important to introduce concepts surrounding the technology starting in elementary school, including how to — and how not — to use it, as well as how it works so students gain an understanding of why the algorithms behave the way they do.
Code.org, a nonprofit dedicated to expanding computer science education, especially to underrepresented populations, is still exploring just how early an understanding of AI should be built, said Karim Meghji, chief product officer. The organization started out aiming its offerings at high school students, and then it filtered down to middle schoolers and, occasionally, the later grades in elementary school.
“I can’t say we have the conviction that you should teach [AI] to kindergartners, but I do think you should start early — not just how to use it, but how the actual technology under the hood is operating,” he said.
The Computer Science Teachers Association agrees that it’s important to teach students how AI works — and CSTA does advocate for at least some instruction as early as kindergarten, said Jake Baskin, executive director, who encourages educators to check out CSTA’s “AI Learning Priorities” report.
As early as kindergarten, “students are starting to recognize patterns,” Baskin said. “Machines are looking at and recognizing patterns to make their decisions. … By 1st grade, students are investigating how patterns can be used by people.”
Building that early foundation creates the potential for 5th graders, for example, to be able to understand how facial recognition can identify some people better than others, based on the biases built into models, Baskin said. By high school, students could be working on developing an AI model for a chosen task using appropriate data and tools.
“What really is clear to me, from this disruption, is that foundational computer science curriculum is more important than ever,” Baskin said. “We’re watching AI dramatically increase the ways computation and computers will be part of everyone’s life going forward. Having a foundational understanding of how these work and how to have agency in creating and interacting with AI is about how we prepare every student to be an engaged citizen, not just preparation for a specific job.”
Meghji points out that if students are taught how to use AI without an understanding of how it works, they’re not going to know why it makes an error, for example.
“AI is primarily mathematics and algorithms with data,” he said. “We’re unpacking the ‘magical’ thing that happens. We’re demystifying it: AI is not magic, it’s operating in a very deterministic way.”
As students gain greater experience with AI, they should also be taught how to evaluate its responses, bringing the “human in the loop,” Meghji said. By middle and high school, students should be prompted to think more deeply about the technology’s societal impacts.
“You’ve got to get into discussions about bias and misinformation, and talk about the ‘why,’” he said, “so students aren’t just users but creators, cognizant of the impacts.”
Baskin also sees a “deep opportunity” for students to consider the societal impacts of AI and how to ethically design AI systems. “There’s an incredibly positive opportunity and significant negative impacts that we need to be aware of and mitigate where possible,” he said.
It’s essential that students understand how these systems work, “so they can make informed decisions as to how and when it is appropriate to do so. There’s no clear [ethical] line of, ‘This is OK, and this isn’t.’ It’s something that’s going to be developed communally.”
Ultimately, students need to understand that AI will play a major role in their lives, Meghji said.
“We live in a digital world,” he said. “Understanding the digital world enables students to become better consumers and citizens.”
