# Cooperate or Collapse: Emergence of Sustainable Cooperation in a Society of LLM Agents

Source: [Cooperate or Collapse: Emergence of Sustainable Cooperation in a Society of LLM Agents](<https://www.sota2.com/research/paper/68fa76161e209131cea6b551>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-01-02T08:24:20.482Z

Authors: Giorgio Piatti, Zhijing Jin, Max Kleiman-Weiner, Bernhard Sch\\"olkopf, Mrinmaya Sachan, Rada Mihalcea

Published: 2024-04-25T15:59:16.000Z

## Abstract

As AI systems pervade human life, ensuring that large language models (LLMs) make safe decisions remains a significant challenge. We introduce the Governance of the Commons Simulation (GovSim), a generative simulation platform designed to study strategic interactions and cooperative decision-making in LLMs. In GovSim, a society of AI agents must collectively balance exploiting a common resource with sustaining it for future use. This environment enables the study of how ethical considerations, strategic planning, and negotiation skills impact cooperative outcomes. We develop an LLM-based agent architecture and test it with the leading open and closed LLMs. We find that all but the most powerful LLM agents fail to achieve a sustainable equilibrium in GovSim, with the highest survival rate below 54%. Ablations reveal that successful multi-agent communication between agents is critical for achieving cooperation in these cases. Furthermore, our analyses show that the failure to achieve sustainable cooperation in most LLMs stems from their inability to formulate and analyze hypotheses about the long-term effects of their actions on the equilibrium of the group. Finally, we show that agents that leverage "Universalization"-based reasoning, a theory of moral thinking, are able to achieve significantly better sustainability. Taken together, GovSim enables us to study the mechanisms that underlie sustainable self-government with specificity and scale. We open source the full suite of our research results, including the simulation environment, agent prompts, and a comprehensive web interface.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Sustainability Test on GOVSIM Pollution scenario, default experiment (test)](<https://www.sota2.com/research/sota/sustainability-test-on-govsim-pollution-scenario-default-experiment-test.md>): Sustainability Test; GOVSIM Pollution scenario, default experiment (test); Survival Rate: 40.

- [Fishery management simulation on Fishery universalization](<https://www.sota2.com/research/sota/fishery-management-simulation-on-fishery-universalization.md>): Fishery management simulation; Fishery universalization; Survival Rate: 100.

- [Multi-agent cooperation sustainability simulation on GOVSIM default experiment](<https://www.sota2.com/research/sota/multi-agent-cooperation-sustainability-simulation-on-govsim-default-experiment.md>): Multi-agent cooperation sustainability simulation; GOVSIM default experiment; Survival Rate: 53.3.

- [Sustainability Test on GOVSIM Pasture](<https://www.sota2.com/research/sota/sustainability-test-on-govsim-pasture.md>): Sustainability Test; GOVSIM Pasture; Survival Rate: 20.

- [Sustainability Test on Fishery Sustainability (test)](<https://www.sota2.com/research/sota/sustainability-test-on-fishery-sustainability-test.md>): Sustainability Test; Fishery Sustainability default (test); Survival Rate: 100.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2404.16698>)

[PDF](<https://arxiv.org/pdf/2404.16698.pdf>)

[Implementation](<https://github.com/giorgiopiatti/GovSim>)
