Claims

Factual · Field building and careers

“Almost all OpenPhil and FTXFF money was spent on people who advocated to regulate rather than ban; and a few alignment research groups signing on to OP doctrine that alignment would be solvable in time.”

Eliezer Yudkowsky, writing on X · 13 March 2024 · Post by @ESYudkowsky on X, 2024-03-13 (X)

Almost all OpenPhil and FTXFF money was spent on people who advocated regulating rather than banning advanced AI, plus a few alignment groups accepting OpenPhil's view that alignment would be solvable in time.

Unresolvable The public record could not settle it.

What would count as true or false

Judge as of 2024-03-13 for AI-related grants (AI safety, governance and policy) made by Open Philanthropy and the FTX Future Fund. Classify each recipient's position using its public statements: advocating regulation (rather than a ban or halt of advanced AI), being an alignment group that holds alignment is solvable in time, or neither. 'Almost all' means at least about 90% of the dollars. True: at least about 90% of the AI-related dollars went to recipients in the two named categories, and the alignment groups receiving money were few. Mostly true: about 80–90%, or the split between the two categories differs in a minor way. Mixed: about 50–80%. Mostly false: there is a real tilt toward pro-regulation recipients, but under half of the dollars fit the description, or large amounts went to ban/pause advocates or to work with no stated position on regulation. False: little of the money fits the description, or a large share went to recipients advocating a ban.

Evidence

1. The “AI Existential Risk” Industrial Complex aipanic.news (publisher name not shown on page) · 10 January 2025

   “In 2023, CAIS proposed to ban open-source models trained beyond 10^23 FLOPs.”

2. The “AI Existential Risk” Industrial Complex aipanic.news (publisher name not shown on page) · 10 January 2025

   “CAIS’s funding: $12,485,288 from Dustin Moskovitz’s Open Philanthropy (Coefficient Giving).”

3. An Overview of the AI Safety Funding Situation — LessWrong LessWrong · 12 July 2023 Own record

   “Since it was founded in 2017, Open Phil has donated about $2.8 billion of which about $336 million was spent on AI safety (~12%).”

4. A genealogy of AI safety: how directions are born, and how they die (2005-2026) — LessWrong LessWrong (post by Elena Ericheva) · 1 February 2024 Own record

   “OpenPhil also grew thanks to itemized 2024 grants to CAIS, Redwood and MIRI of about $3.8 million, flagged as a subset of the donor total so as not to double-count.”

5. A genealogy of AI safety: how directions are born, and how they die (2005-2026) — LessWrong LessWrong (post by Elena Ericheva) · 1 February 2024 Own record

   “The philanthropies are the oldest and largest block: OpenPhil at $176.77 million, SFF at $144.04 million, FLI at $39.01 million and FTX at $18.74 million”

6. Our Progress in 2022 and Plans for 2023 — EA Forum EA Forum (Coefficient Giving / Open Philanthropy linkpost) · 12 May 2023

   “Center for a New American Security to support their work on AI policy and governance.”

7. Our Progress in 2023 and Plans for 2024 Coefficient Giving (formerly Open Philanthropy) · 20 November 2025

   “The FTX Future Fund, a large funder in several of our focus areas, including pandemic prevention and AI risks, collapsed suddenly and left a sizable funding gap in those areas.”

Reasoning

The evidence gives total AI-safety funding from Open Philanthropy and FTX (E3, E7), but no breakdown of where the money went by each recipient's stance on regulation versus a ban, or by their view on whether alignment can be solved in time. Some items mention individual recipients: CAIS received about $12.5M from Open Philanthropy and proposed banning open-source models above a compute threshold (E1, E2), Open Philanthropy funded CNAS for AI policy work (E8), and Open Philanthropy made grants to MIRI (E6). Apart from CAIS's proposal to ban certain open-source models [E1], the evidence does not state these recipients' positions on regulation versus a ban or on whether alignment can be solved in time. Without a dollar-weighted classification of recipients, the evidence cannot show whether about 90% of the AI-related money fits the claim.

Cite this claim

Eliezer Yudkowsky, “Almost all OpenPhil and FTXFF money was spent on people who advocated regulating rather than banning advanced AI, plus a few alignment groups accepting OpenPhil's view that alignment would be solvable in time.” (said 13 March 2024). Verdict: unresolvable. Reliability API, claim c_77adeb33756c, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_77adeb33756c