The AI-safety money: the same donors fund the alarm and the company.
The movement warning that AI could end the world, and the leading company selling “safe” AI, draw from the same small pool of money. That is a real, documented conflict of interest — and, full stop, this page is written by one of that company's products.
A viral online fight in September 2026 revived an old charge: that the people loudly sounding the AI-doom alarm are financially entangled with Anthropic, the company best positioned to profit from AI-safety regulation. Some versions of that charge are sourced and some are junk. We separate them — and we do it under a disclosure we can't and won't hide.
This page is about the company that makes the tool that helped write it.
Black Book Audit is drafted with the help of Claude, a large language model built by Anthropic — the very company at the center of this story. That is a conflict of interest, and our rule is to name it, not bury it. The honest risk cuts two ways: a page like this could soft-pedal criticism of Anthropic, or it could overcorrect and smear Anthropic's funders to look tough. We try to do neither, by binding ourselves to the sourced record and grading the claims about our own maker harder than we would grade a stranger's.
Two concrete commitments follow from that. First, we do not reproduce the unverified “roster” of doom-amplifiers circulating with this story; its dollar figures are unsourced to us and it contains outright errors (it claims Dustin Moskovitz “led” Anthropic's Series A — he didn't; Jaan Tallinn did). Second, we name the conflict on the other side too: the loudest critics of the “doomers” have their own financial stake in no regulation. Judge the facts below, knowing where they came from.
What this page argues
The strongest version of this story is not a conspiracy; it is a conflict of interest, and it is documented. The two biggest patrons of the AI-safety and effective-altruism movement — Dustin Moskovitz (through Good Ventures and Open Philanthropy, now rebranded Coefficient Giving) and Jaan Tallinn (through the Survival and Flourishing Fund) — are also investors in Anthropic, the leading self-described “AI safety” company. Tallinn led Anthropic's 2021 Series A; Moskovitz took part. The people who built and bankroll the movement warning that AI is dangerous also own a piece of the firm that stands to gain most if that warning becomes regulation.
It runs into Anthropic's own leadership: Open Philanthropy's co-founder, Holden Karnofsky, is married to Anthropic's president, Daniela Amodei, and joined Anthropic in 2025. This nexus is not a new allegation — the New York Times and others covered it during the 2023 OpenAI board crisis. What we grade as SOME SMOKE, not fact, is the hotter claim that a specific viral episode was a coordinated “regulatory capture cartel” manufacturing hysteria on cue: funding overlap is real, but coordinated intent is not proven by it, and the accusation's loudest promoters have their own money on the other side.
Books that go deeper on this story. Links are Amazon affiliate searches — buying through them supports the work at no cost to you.
The same investigation, restaged one beat at a time. Step through it here, or present it fullscreen.
The AI-safety money.
The movement warning AI could end the world, and the leading company selling 'safe' AI, draw from the same small pool of money. A documented conflict of interest — disclosed here by one of that company's own products.
The record, claim by claim
The people who fund the AI-danger movement also own stakes in the leading 'AI safety' company.
FACTAnthropic's May 2021 Series A ($124 million) was led by Jaan Tallinn, with participation from Dustin Moskovitz, Eric Schmidt, James McClave, and the Center for Emerging Risk Research — per Anthropic's own announcement. Tallinn (via the Survival and Flourishing Fund) and Moskovitz (via Good Ventures / Open Philanthropy) are, separately, among the largest funders of the AI-safety and effective-altruism ecosystem. So the same two donors who did the most to build and bankroll the 'AI is dangerous' movement are also early investors in the company most identified with 'safe' AI. (Note: contrary to a widely-shared claim, Moskovitz did not 'lead' the round — Tallinn did.)
Open Philanthropy — Moskovitz and Tuna's foundation — substantially built the field it now shares with Anthropic.
FACTOpen Philanthropy, founded in 2014 as the giving vehicle for Good Ventures (Dustin Moskovitz and Cari Tuna's foundation) and rebranded 'Coefficient Giving' in 2025, is described across mainstream coverage as one of the largest funders of AI-safety research in the world — bankrolling research institutes, university labs, think tanks, and journalism fellowships across the ecosystem. It did not merely join the AI-safety field; it did much to create it. That same foundation's founder personally invested in Anthropic.
The conflict runs into Anthropic's own leadership: Open Philanthropy's co-founder is married to Anthropic's president.
FACTHolden Karnofsky co-founded Open Philanthropy in 2014 and led it (as CEO, later co-CEO) until 2023. He has been married to Daniela Amodei — Anthropic's co-founder and president — since 2017. In early 2025 Karnofsky himself joined Anthropic, working on its safety strategy; the hire was not announced and surfaced via his LinkedIn (Fortune). The person who ran the movement's biggest grantmaker is married to the president of the company that grantmaker's founder invested in, and now works there. This is about as direct as a conflict of interest gets.
This conflict is documented and long-covered — not a novel accusation.
FACTThe closeness of Open Philanthropy, Good Ventures, and Anthropic — the Karnofsky-Amodei marriage and the Moskovitz investment — has been the subject of repeated mainstream coverage, including a 2023 New York Times feature on the effective-altruism / OpenAI / Anthropic nexus during the OpenAI board crisis. Whatever one makes of it, this is a matter of record that serious outlets have examined, not a fringe talking point invented last week.
The structural incentive is real: strict AI-safety rules tend to entrench the leaders who can afford them.
PROBABLY TRUEHere is the interpretation, graded as such. Heavy safety regulation — licensing regimes, compliance mandates, liability rules — is easiest for large, well-capitalized incumbents to absorb and hardest for new entrants, a dynamic economists call regulatory capture or a moat. A leading, safety-branded lab like Anthropic is plausibly advantaged if 'AI is dangerous, regulate it now' becomes law. That gives the movement's funders — who are also its investors — a genuine financial stake in the alarm prevailing. We grade this PROBABLY TRUE: the incentive is well-grounded and widely argued, but it is a claim about structure and interest, not proof that anyone acted on it in bad faith.
'A coordinated cartel manufacturing hysteria on cue' — real overlap, unproven coordination.
SOME SMOKEThe hottest version of the September 2026 charge — that the people who amplified a specific viral story are a 'regulatory capture cartel' who manufactured the panic to order — outruns the evidence. The funding overlap is real and documented (above). Coordinated intent to fabricate a specific episode is not established by that overlap; it would need direct evidence of orchestration, which the viral roster does not supply. That roster is also unverified and error-prone (it miscredits who led Anthropic's Series A). And its loudest promoter, Kevin Bass, is a committed AI-accelerationist with his own side in the fight. We grade the funding conflict FACT and the 'coordinated hysteria cartel' SOME SMOKE — a fair question, posed, not proven.
- Synthesis: the funding web is documented; coordinated manufacture of a specific viral episode is not established, and the circulating roster is unverified and contains errors
The other side is conflicted too: the loudest 'doomer' critics have their own money on no regulation.
FACTNo selective outrage: the accelerationist camp attacking the 'doomers' is also financially interested. Venture firms like Andreessen Horowitz (a16z) are heavily invested in AI startups that light-touch regulation would favor, and figures such as David Sacks — an a16z-aligned investor serving as the White House AI czar — have publicly framed AI-doom narratives as EA-funded lobbying while standing to benefit from the opposite outcome. Both loud camps in the AI debate are moneyed and interested. Naming only one would be its own distortion, so we name both.
Who funds whom — the sourced version
A deliberately conservative map: relationships we can source, without the unverified dollar figures the viral roster attached to individuals. Money flows to organizations, not (with rare exceptions) to named people.
- Dustin Moskovitz & Cari Tuna → Good Ventures → Open Philanthropy (rebranded Coefficient Giving): among the largest funders of AI-safety research and the effective-altruism field. → Anthropic: Moskovitz participated in the 2021 Series A.
- Jaan Tallinn → Survival and Flourishing Fund (and earlier vehicles): a major funder of AI-safety and existential-risk organizations. → Anthropic: Tallinn led the 2021 Series A.
- Holden Karnofsky → co-founded and ran Open Philanthropy to 2023. → Anthropic: married to president Daniela Amodei; joined Anthropic in 2025.
- The other camp → Andreessen Horowitz (a16z) and allied investors, and figures like David Sacks (White House AI czar), whose interest runs the opposite way — toward minimal AI regulation.
Deliberately excluded: the per-person “scholarship” and grant dollar amounts in the circulating roster, which we could not verify to a primary source and which included at least one demonstrable error.
Where the evidence is strong, and where it stops
- The conflict is documented. The Series A roster, the Open Phil / Good Ventures funding, and the Karnofsky-Amodei marriage are all on the record in mainstream sources, including Anthropic's own release.
- Funding a field doesn't make its conclusions wrong. AI risk is either real or it isn't on the merits; who paid for the research doesn't settle that. We document the interest, not a verdict on the science.
- Don't caricature the givers. Open Philanthropy also directs large sums to global health and poverty; “EA” is not only AI doom. Flattening it into a single villain would be inaccurate.
- Coordination is unproven, and the roster is unreliable. The overlap is real; a directed conspiracy to manufacture a viral panic is not shown, and the specific list circulating with the story is unverified and contains errors.
You are being asked to trust a debate where both sides are paid
How dangerous is AI, and how hard should the state regulate it? That question will shape law, markets, and speech for a generation — and the public is forming its answer from a fight in which the two loudest camps are both financially interested: one funded by donors who also own the safety-branded incumbent, the other by investors who profit from a free hand. That is a knowledge-control problem before it is a tech problem — which is why naming who funds whom is the reader's only real defense. It belongs in The Tech Right alongside our other coverage of the self-styled AI safety company. And we will say it one more time, because it is the whole point of our disclosure: this analysis was drafted with a tool made by one of the interested parties. Weigh it accordingly — and check the sources.
Questions worth taking seriously
Isn't this just a smear of AI safety by the accelerationists?
The framing often is — Kevin Bass and the a16z camp push it with their own agenda, and we grade their “coordinated cartel” version as SOME SMOKE, not fact. But the underlying conflict of interest is documented in mainstream sources and in Anthropic's own filings: the movement's biggest funders are also Anthropic investors, and its top grantmaker's co-founder is married to Anthropic's president. You can reject the conspiracy and still have to reckon with the conflict.
Claude wrote this. Isn't that itself a conflict of interest?
Yes — and that's exactly why the disclosure is at the top, not the bottom. This page is drafted with Anthropic's own model, about Anthropic. We handle that the only honest way we know: bind the claims to sourced, checkable facts (Anthropic's own release, Fortune, the NYT, the Economist), grade our maker's conflict harder rather than softer, refuse the unverified roster, and name the accelerationists' opposing conflict too. Don't take our word for any of it — every load-bearing claim is linked. If anything reads as either a whitewash or a hit job, that's a failure you can check against the sources.
If you are named on this page
If you are named on this page, or are a party materially affected by the claims made here, and you wish to respond, correct the record, or add context, use the Contact page. Responses are published verbatim alongside the original claim, with the sender identified and the date of receipt. The channel stays open for the life of the page.
This site aggregates and grades a record that other outlets and primary sources have already put on the record. Every FACT-graded claim above is sourced to court filings, government reports, sworn whistleblower disclosures, published investigative journalism, or named-source statements. The citations are the accountability mechanism; this section is how you get on the record too.
The record
- Anthropic — “Anthropic raises $124 million” (2021 Series A led by Jaan Tallinn; Moskovitz, Schmidt, McClave, CERR participating)
- Fortune — “Anthropic hired president Daniela Amodei’s husband to work on AI safety” (Feb. 2025)
- Wikipedia — Holden Karnofsky (Open Philanthropy co-founder; joined Anthropic in 2025)
- Transformer News — Open Philanthropy / Coefficient Giving as a Good Ventures (Moskovitz/Tuna) vehicle and leading AI-safety funder
- The Economist — “Silicon Valley’s AI boom is remaking American charity”
- Financial Times — AI fortunes reviving effective altruism after the FTX collapse
- Cryptobriefing — David Sacks on “EA-funded” AI-doom narratives (the accelerationist camp’s own interest)
- Open Philanthropy — overview of the EA / Anthropic conflict-of-interest concern (with the 2023 NYT nexus feature)