Steven Pinker and Beff Jezos sign CAIS's statement that Anthropic is not a safe AI company
AI safety researchers, PauseAI and e/acc's founder signed the same paragraph, and Dan Hendrycks says other groups were afraid to.
Part of a running story: Anthropic ipo (3 parts)
On November 18, 2026, the Center for AI Safety published a short statement that says "Anthropic is not a safe AI company." It went up at misanthropicstatement.com as Anthropic heads into its initial public offering, signed by people who'd normally be on opposite sides of the AI argument.
CAIS announced it on X at 10:04 am Eastern. When we read the page at 4:09 pm, it listed 92 names, running from Steven Pinker and Lawrence Lessig to George Hotz and Guillaume Verdon, who posts as Beff Jezos.
- Who
- Center for AI Safety and a coalition of signatories
- Published
- November 18, 2026, at misanthropicstatement.com
- Aimed at
- Anthropic, ahead of its initial public offering
- Signers include
- Steven Pinker, Thomas Wolf, Igor Babuschkin and Max Tegmark
What the statement says
It's one paragraph. It says Anthropic warns that recursively self-improving AI could concentrate power and escape human control, then races to build it first, and that the company has quietly weakened its safety pledges.

Anthropic is not a safe AI company, and we must reject its reckless attempt to amass power in the name of the greater good.
The CAIS release lays out its case. In February, it says, Anthropic weakened its Responsible Scaling Policy, dropping its unilateral commitments to pause altogether. It also accuses the company of lobbying against state AI bills such as SB 1047 and the RAISE Act, and of pitching recursive self-improvement to public-market investors as a core competitive advantage.
Those are CAIS's descriptions. We haven't seen Anthropic answer them in public yet.
Why the signers are the story
The list is strange company. Verdon founded the e/acc movement, and CAIS's release counts accelerationists in the coalition alongside AI safety researchers and ethicists. The heads of PauseAI Global and PauseAI US signed too. So did Thomas Wolf of Hugging Face and Igor Babuschkin, an xAI co-founder.
Dan Hendrycks, who runs CAIS, said in the release that people in the coalition "disagree on almost everything," but agree on the danger of a company convinced its own dominance is the greater good. A minute after CAIS's post, he went further on X.
Numerous AI safety groups told us privately they agreed with this statement, yet felt unable to sign without risking relationships or future funding. Anthropic has almost, but not entirely, captured the ecosystem meant to hold it accountable.
He didn't name the groups. About 50 minutes later he posted that he'd stopped advising AI companies, and would stay independent until there's a meaningful slowdown or moratorium.
What people said back
Stephen Casper, an assistant professor at the Harvard Kennedy School who signed, went harder than the statement itself. On X he accused Anthropic of "recklessness and serial dishonesty" and said that put investors and humanity at risk.
Nathan Lambert, who isn't on the list, made a narrower point that a lot of readers will share.
I think it’s good for the ai safety community to strongly say anthropic isn’t the only one to listen to or trust.
- Safety CAIS posts the statement
- Hendrycks Hendrycks on funding
- Casper) Casper signs
- Hendrycks Hendrycks stops advising
Nov 18, Eastern time, read off each post on X.
What it means for people buying the shares
The timing is aimed at investors. Anthropic has applied to list its Class A stock on the Nasdaq Global Select Market as ANT, and its roadshow announcement puts the expected price between $70.00 and $78.00 a share for 825,000,000 shares. CAIS's argument is that the safety branding is part of what's being sold to those buyers.
Rep. Pat Ryan posted the same morning that once Anthropic is public the pressure to deliver profits will be immediate and immense. He didn't mention the statement, but he's asking the same question about money and caution.
That's the part that reaches past the AI industry. Once the stock lists, anyone who buys a share owns a piece of the company the statement is about.The roadshow announcement says the offer can only be made through a prospectus, under a registration statement filed with the SEC, so whatever Anthropic tells buyers about safety, it tells them in writing.
The statement is still taking signatures on its page. Anthropic's roadshow announcement doesn't give a pricing date.
More on Anthropic
- Anthropic cuts its bug disclosure deadline to 30 days, saying another AI will find the bugNovember 16, 2026
- Anthropic says Mythos got Claude running on AMD chips in three days, almost unaidedNovember 12, 2026
- Anthropic launches Claude Sheets and Whiteboards, but Claude Meetings is still coming soonNovember 10, 2026
- AI Futures Project now expects a superhuman AI researcher in 2027, citing METR's Anthropic reportNovember 10, 2026