Humanlike AI Can Strengthen Women’s Belief in Sexist Stereotypes

Can interactions with humanlike AI strengthen gender stereotype beliefs among people from vulnerable groups? AI models are frequently repeating gender stereotypes, and anthropomorphic features have been shown to increase individuals’ perceptions of AI’s trustworthiness in other settings. Although traditional social cognition research suggests that stereotypes lack malleability, these factors raise the concern that anthropomorphic AI chatbots may strengthen stereotypes in individuals who are vulnerable to believing them. Consistent with this prediction, we report results from four experiments on U.S. adults (N = 2,774) showing that politically conservative women believed the gender-math stereotype in a chatbot's response to be more accurate when the chatbot had lifelike features. The effect was mediated by perceived anthropomorphism (specifically, mind perception) and trustworthiness, and we ruled out an alternative cognitive mechanism for this effect. Neither liberal women nor conservative men showed this effect for the gender-math stereotype; however, our final experiment showed that liberal women were indirectly influenced when the chatbot responded with a different gender stereotype that liberals are more inclined to believe. In a formal model, we speculate that social identity threat and ideological predisposition may converge, making women more susceptible to believing ideology-aligned gender stereotypes when anthropomorphic AI chatbots assert them. Our research shows that stereotype beliefs can be shifted by brief interactions with AI on an individual level, and we argue that this effect could be prevented by socially-conscious AI developers who de-anthropomorphize AI applications. Future research should test whether this effect occurs for other identity groups with stereotypes they are predisposed to believe.