The article by Ruqing Xu discusses the design of an algorithm that aids in decision making. It contemplates whether to act directly based on the algorithm’s prediction or delegate the decision to an agent with private information. The study finds that delegation is optimal only if the principal would make the same decision as the agent. The most informative algorithm may not always be the best, and the optimal algorithm may provide more information about one state and restrict information about the other. The study predicts the underperformance of human-machine collaborations if no measures are taken to mitigate common preference misalignment between algorithms and human decision-makers.
Publication date: 14 Feb 2024
Project Page: https://arxiv.org/abs/2402.09384v1
Paper: https://arxiv.org/pdf/2402.09384
