Social stereotypes are prevalent and consequential; they can result from segregated societies. Why are certain groups stratified into particular positions? In this talk, Dr. Bai will propose a simple yet powerful psychological mechanism: a seemingly intelligent, self-interested exploration at the individual level that cascades into collateral damage of segregated structures at the collective level. The key insight considers stereotype formation in sequential decisions, where exploring new options is costly. Using stereotypes of immigrants in the United States as a motivating example, Dr. Bai will show how computational cognitive modeling and large-scale online behavioral experiments can inform this proposal. This idea illustrates an insidious psychological possibility for well-intentioned and attentive decision-makers to create biased societies. Dr. Bai will discuss promising directions and critical caveats for diversity science and fairness policy in human and artificial intelligence systems.