Welfare Diplomacy: Benchmarking Language Model Cooperation

TL;DR

This paper introduces Welfare Diplomacy, a new benchmark game for evaluating AI cooperation that balances military and domestic welfare, enabling better assessment and training of cooperative multi-agent systems.

Contribution

It proposes Welfare Diplomacy as a novel benchmark, implements it with open-source tools, and evaluates baseline language model agents on social welfare and exploitability.

Findings

High social welfare achieved by state-of-the-art models

Baseline agents are exploitable despite high welfare

Welfare Diplomacy provides clearer cooperation assessment

Abstract

The growing capabilities and increasingly widespread deployment of AI systems necessitate robust benchmarks for measuring their cooperative capabilities. Unfortunately, most multi-agent benchmarks are either zero-sum or purely cooperative, providing limited opportunities for such measurements. We introduce a general-sum variant of the zero-sum board game Diplomacy -- called Welfare Diplomacy -- in which players must balance investing in military conquest and domestic welfare. We argue that Welfare Diplomacy facilitates both a clearer assessment of and stronger training incentives for cooperative capabilities. Our contributions are: (1) proposing the Welfare Diplomacy rules and implementing them via an open-source Diplomacy engine; (2) constructing baseline agents using zero-shot prompted language models; and (3) conducting experiments where we find that baselines using state-of-the-art models attain high social welfare but are exploitable. Our work aims to promote societal safety by aiding researchers in developing and assessing multi-agent AI systems. Code to evaluate Welfare Diplomacy and reproduce our experiments is available at https://github.com/mukobi/welfare-diplomacy.