Human Exploration in Complex Problem-Solving Tasks: More Effortful Interaction Leads to Higher Efficiency

Exploration, a cornerstone of the human ability to solve novel problems, is a complex process. Most studies on human exploration used overly simple tasks that isolate variables but poorly reflect problems humans evolved to solve—limiting the generalizability of the results. To address this limitation, we introduce the Lockbox paradigm, a novel, ecologically valid, and challenging task that requires active exploration and physical interaction. Data from 263 participants interacting with the Lockbox across three different interaction modalities of varying interaction costs, reveal a remarkable ability to adapt and solve problems efficiently in complex scenarios. By comparing the interaction modalities, we demonstrate the critical role of cost variations, such as physical and temporal costs, in driving attentiveness and shaping exploration strategies. These findings provide important insights into human exploration strategies, with potential applications in fields such as robotics and artificial intelligence.