Commitment To Cooperation With Self-Negotiated Contracts

2026-07-23

AI Summary

In multi-agent environments, AI agents struggle to establish credible commitments due to the upfront costs of cooperation and delayed benefits, leading to inefficient collaboration. This work proposes a self-negotiated contract mechanism inspired by legal agreements, which uniquely integrates formal contracts with natural language to enable large language model–based agents to autonomously reach executable, credible commitments in spatiotemporal games without external enforcement. The framework supports diverse contract representations—ranging from compilable code to natural language—and jointly trains navigation and negotiation capabilities within a CT (cooperative transaction) game environment. Experimental results demonstrate that the proposed mechanism substantially outperforms conventional transaction methods, significantly improving both cooperation success rates and overall collective utility.

Abstract

As AI agents operate with increasing autonomy in a multi-agent world, they will need to learn to cooperate with other agents and with humans to generate mutual benefits. However, cooperation is a challenge because the costs of cooperation are often incurred early on, but the benefits are only realized later, creating an incentive to defect. How can AI agents cooperate with commitment? Here, we draw on inspiration from legal institutions and contracting that human societies have used to solve principal-agent problems of this kind. Contracts provide observable representations of agreements that enable credible commitments through the enforcement of terms. We study the role of contract-based cooperation using LLM-based agents in \CT, a spatial-temporal game that combines bargaining with navigation towards a goal. We study a suite of contract representations that range from formal contracts that compile to code to natural contracts that require reinterpretation. We evaluate agents with a range of LLM backbones using different sizes and providers. We find that self-negotiated contracts can improve cooperative outcomes beyond what is possible with regular trading.

Problem

• cooperation
• commitment
• multi-agent systems
• principal-agent problem
• contracts

Innovation

• self-negotiated contracts
• multi-agent cooperation
• large language models
• contract-based commitment
• spatial-temporal game