We conduct interdisciplinary AI research at the intersection of safety, governance, and the humanities and social sciences. Our nonprofit enables fast, flexible collaboration between AI researchers and domain experts across institutions, disciplines, and sectors — filling gaps that traditional academic structures can't. Founded by Katherine Elkins and Jon Chun, who co-created the world's first human-centered AI curriculum in 2016, our founders lead the Modern Language Association team at the NIST U.S. AI Safety Institute Consortium (CAISI) and the Schmidt Sciences HAVI Archival Intelligence project.
LLM red/blue team testing, ethical auditing, multi-agent system evaluation, and comparative AI regulation. Our research contributes to AI safety and governance conversations across standards bodies, public-interest AI organizations, and international cultural-policy initiatives.
An umbrella for distributed teams of diverse experts to secure funding and collaborate on AI research across geographic, institutional, and disciplinary boundaries.
All research tools and methodologies are open-source. Student research from the curriculum we co-created has been downloaded more than 100,000 times by more than 4,000 institutions worldwide.