RCBSF: A Multi-Agent Framework for Automated Contract Revision via Stackelberg Game

2026-04-12

AI Summary

This work addresses the challenges of hallucination and insufficient behavioral constraints in large language models (LLMs) when performing automated contract revision. To mitigate these issues, the authors propose a bilevel multi-agent framework that formalizes the revision process as a non-cooperative Stackelberg game, comprising a global normative agent, a constrained revision agent, and a local verification agent. A risk budgeting mechanism is introduced to enable iterative optimization under explicit risk constraints. This approach represents the first application of a bilevel Stackelberg game to contract revision with rigorous theoretical guarantees of convergence. Experimental results on a unified benchmark demonstrate that the proposed method achieves an average risk resolution rate of 84.21%, significantly outperforming existing iterative baselines while simultaneously improving utility, safety, and token efficiency.

Technology Category

• Multiagent Systems: Mechanism Design
• Natural Language Processing: Safety and Robustness
• Game Theory and Economic Paradigms: Mechanism Design

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search

Abstract

Despite the widespread adoption of Large Language Models (LLMs) in Legal AI, their utility for automated contract revision remains impeded by hallucinated safety and a lack of rigorous behavioral constraints. To address these limitations, we propose the Risk-Constrained Bilevel Stackelberg Framework (RCBSF), which formulates revision as a non-cooperative Stackelberg game. RCBSF establishes a hierarchical Leader Follower structure where a Global Prescriptive Agent (GPA) imposes risk budgets upon a follower system constituted by a Constrained Revision Agent (CRA) and a Local Verification Agent (LVA) to iteratively optimize output. We provide theoretical guarantees that this bilevel formulation converges to an equilibrium yielding strictly superior utility over unguided configurations. Empirical validation on a unified benchmark demonstrates that RCBSF achieves state-of-the-art performance, surpassing iterative baselines with an average Risk Resolution Rate (RRR) of 84.21\% while enhancing token efficiency. Our code is available at https://github.com/xjiacs/RCBSF .

Problem

• automated contract revision
• hallucination
• behavioral constraints
• safety
• Large Language Models

Innovation

• Stackelberg Game
• Multi-Agent Framework
• Risk-Constrained Optimization
• Automated Contract Revision
• Bilevel Optimization