The MisAnthropic Statement was prompted by a dispute over whether Anthropic’s public image as a safety-focused AI company matched its actions. Announced by the Center for AI Safety (CAIS) on Nov. 18, 2026, ahead of Anthropic’s IPO, it accused the company of warning about the risks of recursively self-improving AI while racing to build it.
1
14 The statement’s claims are allegations, not independent proof that those risks have materialized.
The safety-branding dispute
CAIS framed the statement around an apparent contradiction: Anthropic warns that recursively self-improving AI could concentrate power and escape human control, while, in the signers’ view, pursuing that technology itself.
1 The statement also challenged the relationship between Anthropic’s safety branding and its effort to build influence and power. That is the coalition’s argument; the announcement does not establish that Anthropic’s branding was intended to mislead investors or the public.
Other reporting summarized the signers’ criticism as accusing Anthropic of weakening safety pledges and lobbying for regulation that served its interests.
2 The evidence provided here supports those broad allegations, but it does not verify the exact February changes to Anthropic’s Responsible Scaling Policy or its specific positions on California’s SB 1047 and New York’s RAISE Act. Those details should not be treated as established by the available excerpts.
A dispute over commitments and collective safeguards
The underlying disagreement is not simply whether AI risks matter. It is also about what a company should promise when it believes safety depends on action across the industry. Reporting describes Anthropic as separating unilateral company commitments from broader industry recommendations in a revised policy.
17 That distinction helps explain the debate: critics can view the loss of a company-level commitment as a retreat, while a company may argue that durable safeguards require wider coordination. The supplied sources do not establish Anthropic’s full argument that acting alone could leave it behind less cautious competitors.
CAIS director Dan Hendrycks said some safety groups privately agreed with the statement but were reluctant to sign because of possible risks to relationships or future funding.
27 That is his account of what those groups told him; it does not independently demonstrate that funding or relationships improperly influenced them.
The signers did not represent a single position on AI development. CAIS described the coalition as spanning AI safety, open-source, accelerationist, scientific and legal communities.
16 And not everyone who criticized Anthropic signed: Gizmodo reported that open-weights advocate Nathan Lambert said the statement was too strong for his taste, despite his own criticism of Anthropic.
12 A shared objection to a company therefore should not be read as agreement among signers about AI risk or regulation.
What the signer counts show
The user-supplied premise says more than 100 people signed. The contemporaneous sources provided here do not substantiate that figure: reports recorded 89, 91 or 92 names, while another said “at least 80.”
2
3
7
8 These counts may reflect different snapshots of a changing list, but the evidence here does not confirm a total above 100 or every name in the question.
The IPO was not a safety verdict
Anthropic priced its IPO at $83 per share, above the marketed $70–$78 range. Its official announcement listed 980 million shares in the offering: 956,351,354 sold by Anthropic and 23,648,646 by existing shareholders.
30
34 Reporting put the company’s valuation at about $1.5 trillion.
31
33
The above-range price is evidence of strong demand for the offering, not proof that the statement had no effect, that no investor reconsidered, or that Anthropic’s safety practices are adequate. Nor should the total offering be described as entirely new money for Anthropic: some shares were sold by existing shareholders, according to the company’s announcement.
30
The best-supported explanation for the statement is therefore a credibility challenge issued just before a major public-market milestone. The coalition argued that Anthropic’s safety message conflicted with its approach to self-improving AI, commitments and regulation. The available evidence documents that dispute and the IPO’s strong pricing; it does not resolve the underlying safety claims.