Expectations about other individuals shape many of our social interactions. But what happens when these expectations come about? In this talk, I will suggest that it is intrinsically rewarding for us to successfully use stereotypes and other sources of knowledge to predict other individuals. In a set of pre-registered studies (total N > 1000), participants made multiple choices between rating a stereotype-confirming and rating a stereotype-violating target. We associated each choice option in each trial with a small monetary payoff that we subsequently added to participants' compensation. Participants consistently chose to interact with stereotype-confirming targets, even when making such choices entailed forgoing higher monetary amounts (Cohen’s ds: 0.25 – 0.4 across studies). To provide further support for the subjective value of stereotype-confirmation, two separate groups of participants underwent fMRI scanning (total N=58) while rating expectation-confirming or expectation-violating targets. Subsequently, participants completed the monetary incentive delay task to functionally localize reward-sensitive neural regions. Reward-related regions increased their activity to expectation-confirming relative to expectation-violating information. We found this effect for expectations emanating from stereotypes as well as for expectations emanating from knowledge about famous individuals. Together, these findings indicate that participants subjectively value the successful prediction of social targets. This subjective value may underlie the motivation to confirm our expectations while interacting with – and interpreting information about – other individuals.