Dan Ho, William Benjamin Scott and Luna M. Scott Professor of Law, and co-authors, including Andy Zhang, JD ’26, won the SafeBench Prize for Cybench, a first-of-its-kind framework for specifying cybersecurity tasks and evaluating AI agents on those tasks. Policymakers, model developers, and researchers in the AI and cybersecurity communities are using the tool to assess agent capabilities, mitigate cyber risks, and explore applications for penetration testing. The project is a collaboration through Stanford’s RegLab.

Additionally, in August 2024, Ho received an Outstanding Paper Award at the 62nd Annual Meeting of the Association for Computational Linguistics for his co-authored article “MultiLegalPile: A 689GB Multilingual Legal Corpus.” The paper presents a large-scale, multilingual legal text dataset designed to support natural language processing applications in the legal domain. It describes the dataset’s construction, sources, and potential applications for training AI models on legal texts across multiple languages. SL