John Rawls’ theory of justice as fairness remains a pivotal framework to understand fairness in social contracts. While highly influential, key concepts from Rawls’ framework for comparative justice have proven difficult to realize in practice. In this paper, we present a novel approach that uses Artificial Intelligence (AI) agents for semi-automatic generation of fairer bilateral legal agreements. Our results demonstrate the utility that Large Language Models (LLMs) can achieve in the analysis of human made agreements. While the approach maintains sensitivity to some of the same vulnerabilities that apply to human-authored agreements, we believe that Rawlsian agents represent a novel application of AI to the authorship of legal documents from a position of fundamental fairness. Potential extensions of this work could reduce systematic inequalities in legal access, reduce litigation risk, and ultimately help realize Rawls’ vision of a fairer society.