Reproduce “The art of compensation: how hybrid teams solve collective risk dilemmas” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Inês Terrucha, Elias Fernández Domingos, Francisco C. Santos, Pieter Simoens, Tom Lenaerts
This research uses an evolutionary game theory model to investigate how human cooperation adapts in collective risk dilemmas when AI agents are present. The study identifies a "compensation effect" where humans either reduce their cooperation (exploiting benevolent AI) or increase it (compensating for non-cooperative AI), ultimately finding that cooperative AI tends to increase overall group success rates, often by reducing human engagement.
Key insights
- A "compensation effect" dictates human response to AI: humans reduce cooperation when AI is highly cooperative (exploitation) but increase it when AI is non-cooperative (compensatory boost).
- The influence of AI on human behavior becomes more pronounced and sharply defined as the number of AI agents increases.
- Maintaining a constant expected AI contribution does not lead to equivalent collective outcomes, highlighting the complexity of human-AI interaction dynamics.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.