A coalition of 92 signers says Anthropic is not a safe AI company. The MisAnthropic Statement went up on November 18, 2026, the same day Anthropic priced its IPO at $83 a share. Its core sentence: “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 Center for AI Safety posted it on X at 15:04 UTC, calling the signers “a rare coalition across AI safety, open source, e/acc, science, and law”. By 01:07 UTC on November 19 that post had about 703,000 views.
On this page · 7 sectionsOpen
A statement published on November 18, 2026, the day Anthropic priced its IPO, says Anthropic is not a safe AI company and accuses it of racing to build recursively self-improving AI.
The page listed 92 signers when we read it at 01:08 UTC on November 19, up from 89 cited in a post on X at 22:11 UTC the evening before.
Signers include 2 Nobel laureates, 3 Fields medalists and a Turing Award laureate, alongside Steven Pinker, Thomas Wolf of Hugging Face, xAI co-founder Igor Babuschkin and Beff Jezos.
The statement’s contact addresses are on safe.ai, the Center for AI Safety’s domain, and its executive director Dan Hendrycks signed it. The Center’s post on X passed 700,000 views.
A separate rally calling for a ban on recursive self-improvement is planned outside Anthropic’s offices on November 19, the day its shares are set to start trading.
§ 01What the statement says
The statement is three sentences. The page introduces it by saying Anthropic “uniquely portrays itself as an AI safety company and leverages that reputation to advance its commercial and political ambitions.”
- On building it. “Anthropic warns that recursively self-improving AI could concentrate power and escape human control, yet races to build it first.”
- On the brand. “The company uses safety branding to court investors, employees, and the public”, and the statement says it has “quietly weakened its safety pledges and lobbied for self-serving regulation.”
- The verdict. “Anthropic is not a safe AI company”.
The statement gives no examples or sources for the weakened pledges or the lobbying. Those are the signers’ claims, not findings we checked.
§ 02Who signed
| Signer | Role on the page |
|---|---|
| Steven Pinker | Johnstone Family Professor of Psychology, Harvard University |
| Oliver Hart | Nobel Laureate in Economics; University Professor, Harvard |
| Judea Pearl | Turing Award Laureate; Chancellor’s Professor of Computer Science, UCLA |
| Thomas Wolf | Co-founder and Chief Science Officer, Hugging Face |
| Igor Babuschkin | Co-founder and CEO, River AI; Co-founder, xAI |
| Guillaume Verdon (Beff Jezos) | Founder and CEO, Extropic |
| Max Tegmark | Professor of Physics, MIT; Founder and Chair, Future of Life Institute |
| George Hotz | Founder and CEO, the tiny corp; Founder, comma.ai |
| Dan Hendrycks | Executive Director, Center for AI Safety |
By our count of the page, 2 signers are Nobel laureates, 3 are Fields medalists and one holds the Turing Award. The list also takes in Laurence Tribe and Lawrence Lessig of Harvard Law School, Douglas Hofstadter, Gary Marcus, Go co-creator Rob Pike, former OpenAI researcher Gretchen Krueger, and leaders of PauseAI and ControlAI.
What stands out is the spread. Pause advocates sit beside Beff Jezos, a leading voice of effective accelerationism, and open-source builders sit beside people who want frontier AI slowed. They agree on one target, not on what should replace it.
§ 03How it spread
| Time | Account | What it said | Views |
|---|---|---|---|
| 15:04 | Center for AI Safety | Posted the statement and the signing link | about 703,000 |
| 17:35 | MTS | Named Wolf, Babuschkin, Soares and Jezos as signers | about 332,000 |
| 18:43 | Daniel Kokotajlo | Will attend the rally outside Anthropic’s offices | about 36,000 |
| 21:30 | Tristan Harris | Promoted the same rally | about 9,000 |
| 22:11 | Coin Bureau | Cited 89 signers and quoted the statement | about 29,000 |
§ 04The rally on trading day
A separate event is planned for the morning Anthropic’s shares are set to start trading. Daniel Kokotajlo wrote: “I will be at the Ban RSI rally outside Anthropic offices tomorrow at 11 am.” Tristan Harris, a signer of the statement, promoted the same rally. RSI is recursive self-improvement, the thing the statement says Anthropic is racing toward.
§ 05What Anthropic has said
Nothing on X about the statement as of 01:08 UTC on November 19. Its most recent post was the IPO pricing: “We’ve priced our IPO at $83 per share, with trading expected to begin on Nasdaq tomorrow.” Our record of the IPO has the filing details.
Anthropic has published its own numbers on how much of its research Claude already does, which is the measure the statement’s critics point to. We covered them in Anthropic’s three numbers.
§ 06Where CellCog sits
CellCog runs on Anthropic’s models: Claude Opus 5.5 on Flash and Core, Claude Fable 5.2 on Agent Max and Creative Max. That gives us a stake in Anthropic, so this page reports what the statement and its signers say and takes no side on it.
§ 07Sources
- The MisAnthropic Statement, misanthropicstatement.com, read 01:08 UTC, November 19, 2026
- Center for AI Safety on X, the statement, 15:04 UTC, November 18, 2026
- MTS on X, signers named, 17:35 UTC
- Daniel Kokotajlo on X, the rally, 18:43 UTC
- Tristan Harris on X, the rally, 21:30 UTC
- Coin Bureau on X, 89 signers, 22:11 UTC
- Anthropic on X, IPO pricing, 23:29 UTC, November 18, 2026
Q1Who is behind the MisAnthropic Statement?
The page names no organization in its body, but its contact addresses are on safe.ai, the Center for AI Safety’s domain, the Center posted it on X, and its executive director Dan Hendrycks is a signer.
Q2How many people signed it?
92 as of 01:08 UTC on November 19, 2026, by our count of the names on the page. A post on X at 22:11 UTC on November 18 cited 89.
Q3What does recursive self-improvement mean?
AI systems that help build better AI systems, which then speed up the next round. The statement says Anthropic warns about it yet races to build it first.
Q4Is the statement connected to the Anthropic IPO?
Yes. The page says the coalition endorsed it in light of Anthropic’s public offering. Anthropic priced its IPO at $83 a share on November 18, with trading expected on Nasdaq on November 19.
Q5Does CellCog use Anthropic's models?
Yes. CellCog runs Claude Opus 5.5 on Flash and Core and Claude Fable 5.2 on Agent Max and Creative Max, so we have a stake in Anthropic. This page reports what the statement says and takes no side.
