A new study led by Rajesh Rao, a UW professor in the Paul G. Allen School of Computer Science & Engineering in collaboration with I-LABS suggests that AI can learn cultural values by observing human behavior. This is the same way that children learn. “Kids learn almost by osmosis how people act in a community or culture. The human values they learn are more ‘caught’ than ‘taught’” says I-LABS Co-Director Andrew Meltzoff.
In the study, Researchers had AI systems observe people from two cultural groups playing a video game. The goal of the game, Overcooked, is to work together to cook and deliver as much onion soup as possible. On average, participants in one group behaved more altruistically. The AI assigned to each group learned that group’s degree of altruism, and was able to apply that value to a novel scenario beyond the one they were trained on.
As inspiration, the team looked to previous I-LABS research showing that 19-month-old children raised in Latino and Asian households were more prone to altruism than those from other cultures. “Parents don’t simply train children to do a specific task over and over. Rather, they model or act in the general way they want their children to act. For example, they model sharing and caring towards others,” says Dr. Meltzoff. This new research suggests AI can learn cultural values in the same way.
