FAR AI

FAR.AI (originally the Fund for Alignment Research, now Frontier Alignment Research) is a 501(c)(3) nonprofit research organization founded in July 2022 and incorporated in October 2022. The organization was co-founded by Adam Gleave (CEO), who holds a PhD in artificial intelligence from UC Berkeley under Stuart Russell and previously worked at Google DeepMind, and Karl Berzins (President/COO). FAR.AI's core mission is ensuring advanced AI systems are safe and beneficial for everyone. The organization focuses on research agendas that are too large to be pursued by individual academic or independent researchers but are too early-stage to be of interest to most for-profit organizations. They take bets on a range of promising early-stage agendas and scale up those that prove most successful. The organization's technical research is organized around three primary areas. Their robustness research has uncovered vulnerabilities in superhuman Go AI systems, work that was published in Nature and covered by the Financial Times and MIT Technology Review. Their value alignment work has developed more sample-efficient value learning algorithms. Their model evaluation direction has produced a variety of new black-box and white-box evaluation methods, including red-teaming leading language models. Beyond in-house research, FAR.AI engages in significant field-building activities. The Alignment Workshop series, inaugurated in February 2023, brings together top machine learning researchers from industry, academia, and government to discuss critical AI alignment topics. FAR.Labs, founded in Q3 2023, is a Berkeley-based coworking space hosting 40+ members from various AI safety organizations. The organization also runs a grantmaking program funded by a $12 million grant from Open Philanthropy to support research at academic institutions. FAR.AI has published over 30 research papers, hosted 10+ events reaching over 1,000 attendees, and had its research cited in U.S. Congressional testimony and covered by The New York Times, Nature, MIT Technology Review, Financial Times, and Ars Technica. The organization has been recommended by Founders Pledge as a high-impact giving opportunity. In 2025, FAR.AI secured over $30 million in funding commitments from Coefficient Giving (previously Open Philanthropy), Schmidt Sciences, Survival and Flourishing Fund, the Center for Security and Emerging Technology (CSET), and the AI Safety Fund (AISF). This funding is enabling the organization to grow from 15 to 30+ researchers, launch a technical governance division collaborating with AI Safety Institutes and policymakers, and establish fellowship programs supporting 20+ fellows annually. Prior to incorporation, FAR.AI operated as a fiscally sponsored project under Players Philanthropy Fund.