Why studying children’s minds could help us build better AI

UC Berkeley psychology professor Alison Gopnik studies how children learn. She also studies artificial intelligence systems. That might sound counterintuitive, but as she explains in this video about her work for UC Berkeley News, there’s much overlap between what we are learning about how babies explore the world and how we might create better AI systems.

“Even very young children are the best learners, and they do it in record time and with very little teaching,” says Gopnik. “So the question is, how do they manage to learn as much as they do?”

“Answering that question not only helps us to understand children, but it can help us understand some deep, profound philosophical questions about how it’s possible for any of us to know as much as we do about the world — children, grownups, scientists, and even artificial machines,” she continues.

Gopnik, who was recently awarded the prestigious Rumelhart Prize in Cognitive Science, has studied how children make sense of their world with very little evidence. 

“I study children because there aren’t any aliens,” says Gopnik, who joined UC Berkeley’s faculty in 1988. “If you want to study little creatures with big heads and little bodies who are incredibly brilliant and spend a lot of the time hypnotizing us, the way to do it is to look at 3- and 4-year-olds.”

Gopnik’s video is the first of a new series from Berkeley News called “101 in 101,” which challenges professors and researchers from across the campus to distill the basics of their work, like an introductory 101 course, in just 101 seconds. 

Consider it the briefest lecture you’ll ever enjoy.