The Double-Edged Sword of Behavioral Responses in Strategic Classification

When humans are subject to an algorithmic decision system, they can choose to strategically adjust their behavior accordingly ("game" the system). While a growing line of literature on strategic classification has used game-theoretic modeling to understand and mitigate such gaming, these existing works consider standard models of fully rational agents. In this paper, we propose a model of strategic classification which takes into account behavioral biases in human responses to algorithms. We show how misperceptions of the classifier (specifically, of its feature weights) can lead to different types of discrepancies between biased and rational agents' gaming responses, and identify when behavioral agents over- or under-investment in different features. We also show that strategic agents with behavioral biases can benefit or (perhaps, unexpectedly) harm the firm compared to fully rational strategic agents, highlighting the need to account for human (cognitive) biases when designing AI systems with strategic human in the loop.