The Unjournal evaluates legal research for AI governance using LLMs

The Unjournal, an organization known for evaluating academic research, is expanding into legal scholarship with a focus on high-impact areas like AI governance, animal welfare, and global trade. In a recent update, David Reinstein explains that while the pilot has not yet launched—primarily due to the lack of a senior legal scholar to co-lead—the groundwork is laid. A prototype tool now sources, curates, and rates legal research using reinforcement learning from human feedback (RLHF), incorporating LLM-generated content to assist prioritization. The tool includes input forms for users to suggest and rate work, with potential future incentives for contributions.

Reinstein emphasizes the relevance to US policy on AI safety, ODA (official development assistance), and democracy issues. The project is still actively seeking a legal scholar willing to put their name behind it, calling it a necessary condition to move forward. For now, the prototype demonstrates how AI-driven evaluation can surface underappreciated legal work, potentially reshaping how policymakers and funders allocate attention to legal research with global impact.