This article experiments with processes for using large language models to construct a Restatement of Torts based upon the model’s understanding of leading torts cases. The model’s performance is evaluated by comparing its results with existing provisions from the human-written Restatements of Torts. By tracking the discrepancies between the restatements, we can gain insights into both the machine and human processes for understanding and synthesizing the law. Where the machine and human restatements converge would tend to indicate the reliability of both sources on a particular subject. Where the restatements diverge may reflect the limitations of a language model or human author, meaningful differences in how humans and machines process information, or different underlying values. This analysis has implications for the potential of large language models as tools for legal research and writing, accountability mechanisms for computer and human authorship, the function and authority of Restatements of the Law, the future of human-machine collaboration in legal practice, and the potential for machine learning to reshape the law itself.