Learning Reinforces Curiosity

Human curiosity is often seen as a dynamic process, constantly shifting its focus. However, the principles governing these fluctuations remain debated. Here, we test two competing hypotheses on how past learning experiences shape subsequent curiosity and memory. The first hypothesis, based on the “optimal arousal” theory, proposes that satisfying curiosity leads to its reduction. In contrast, the reinforcement learning perspective suggests that satisfying curiosity strengthens it, leading to greater curiosity in the future. To distinguish between these accounts, we collected information-seeking decisions from 5,831 participants, who chose whether to wait for answers to various questions. We examined how question and answer engagement, as well as information prediction errors, influenced subsequent curiosity. Our findings show that reading satisfying answers increases curiosity compared to reading dissatisfying ones. Crucially, this effect is modulated by semantic similarity—prior learning enhances subsequent curiosity only when new information is related to the previously learned content. These results indicate that curiosity operates as an information-seeking policy learned through reinforcement. We suggest that humans seek information not only to improve future instrumental decisions, but also to learn what to be curious about.