OPEN PHILANTHROPY (COEFFICIENT GIVING)

Institutional Archetype

THE FUNDER OF THE REFEREES — Open Philanthropy, now Coefficient Giving, is the largest single source of money behind the AI-safety field: the evaluators, the think tanks, the fellowships and the career pipelines that the labs, governments and press rely on to judge frontier AI. In 2017 it also paid for a seat at the table of a lab. Its people moved into the labs, the evaluators and government. The file is open because so much of the outside check on frontier AI traces back to one checkbook and one family’s fortune.

Mandate & Origin

“Coefficient Giving is a philanthropic funder and advisor. Since 2014, we’ve directed over $7 billion in grants”; “Our founding and most significant partnership is with Good Ventures, for whom we serve as outsourced foundation staff” (Coefficient Giving). Its own history runs GiveWell Labs (2011–2014), Open Philanthropy (2014–2025), Coefficient Giving (2025–) (Coefficient Giving, Our History). Good Ventures is the foundation of Dustin Moskovitz and Cari Tuna.

Funding & Backers

Good Ventures, plus other donors it advises (Coefficient Giving). Its per-grant pages moved when the name changed and several now redirect; the amounts for individual evaluator grants are in the recheck queue rather than stated here.

Position in the Apparatus

• Since 2015. “Open Philanthropy began supporting AI safety and security work in 2015, when the field was in its early stages. We focused on building institutions and supporting talent pipelines” (Alexander Berger and Emily Oehlsen, Oct 1, 2025).
• A lab, in 2017. “In 2017, Moskovitz’s Open Philanthropy provided a $30 million grant to OpenAI,” which its spokesperson said was “to support work on AI safety, not an equity investment” (POLITICO, Oct 13, 2023). Its co-founder Holden Karnofsky served on OpenAI’s board (Fortune).
• Its people, now inside. Karnofsky joined Anthropic as a member of technical staff in January 2025, working on its Responsible Scaling Policy (Fortune). Ajeya Cotra, who led its technical AI-safety grants, is now at METR. The full map is in The EA to AI-Safety Funding Pipeline and The AI Evaluators.
• Asking for company. In October 2025 it argued “that now is a uniquely high-impact moment for new philanthropic funders to enter the field of AI safety and security” (Coefficient Giving, Oct 2, 2025).

Actions & Leadership Choices

CONDUCT: STRUCTURAL — FUNDS THE LABS’ CRITICS, AND ONCE A LAB. No wrongdoing is alleged. The structural point is concentration: the evaluators that test the labs, many of the researchers who study them and some of the staff who now run their safety policies were funded or employed by one philanthropy, which also funded and sat on the board of the lab whose race the field now argues about. The funder says so itself and asks others to join it, which is the strongest case for the arrangement: without it there might be no outside check at all.

Reach Assessment

Institutional: Over $7 billion in grants since 2014, with AI safety a named priority since 2015. Memetic: Its worldview, catastrophic risk from advanced AI, became the frame through which much of the press and many governments read frontier AI.

Sources: About — Coefficient Giving; Our History — Coefficient Giving; Our Approach to AI Safety and Security — Coefficient Giving (Oct 1, 2025); AI Safety and Security Need More Funders — Coefficient Giving (Oct 2, 2025); POLITICO (Oct 13, 2023); Fortune (Feb 13, 2025).