Joint Scoring Rules: Zero-Sum Competition Avoids Performative Prediction
This addresses a fundamental conflict of interest in decision-making scenarios where predictive agents might manipulate outcomes, offering a solution for principals relying on expert advice.
The paper tackles the problem of expert agents manipulating predictions to influence a principal's decisions, showing that zero-sum competition among multiple agents eliminates this incentive and allows the principal to choose their true preference. Experiments in a toy environment demonstrate that this approach significantly improves predictive accuracy and principal utility while eliminating manipulative behavior.
In a decision-making scenario, a principal could use conditional predictions from an expert agent to inform their choice. However, this approach would introduce a fundamental conflict of interest. An agent optimizing for predictive accuracy is incentivized to manipulate their principal towards more predictable actions, which prevents that principal from being able to deterministically select their true preference. We demonstrate that this impossibility result can be overcome through the joint evaluation of multiple agents. When agents are made to engage in zero-sum competition, their incentive to influence the action taken is eliminated, and the principal can identify and take the action they most prefer. We further prove that this zero-sum setup is unique, efficiently implementable, and applicable under stochastic choice. Experiments in a toy environment demonstrate that training on a zero-sum objective significantly enhances both predictive accuracy and principal utility, and can eliminate previously learned manipulative behavior.