Anthropic is Not a Safe AI Company

November 18, 2026

SAN FRANCISCO, CA — November 18, 2026 Ahead of Anthropic’s upcoming initial public offering, the Center for AI Safety (CAIS) today joined a rare coalition of prominent figures in signing the MisAnthropic Statement.

The statement challenges the disconnect between Anthropic’s public “safety” branding and its reckless attempt to build recursively self-improving AI and consolidate power. 

Published today at misanthropicstatement.com, the statement reads:

Anthropic warns that recursively self-improving AI could concentrate power and escape human control, yet races to build it first. The company uses safety branding to court investors, employees, and the public—even as it has quietly weakened its safety pledges and lobbied for self-serving regulation. 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 coalition brings together AI safety researchers, open-source advocates, AI ethicists, and accelerationists – united by a shared rejection of Anthropic’s effort to position itself to the public and policymakers as the industry’s “good guys.”

Prominent signatories include:

  • Steven Pinker (Johnstone Family Professor of Psychology, Harvard University)
  • Oliver Hart (Nobel Laureate in Economics; University Professor, Harvard University)
  • Judea Pearl (Turing Award Laureate; Chancellor’s Professor of Computer Science, UCLA)
  • Laurence Tribe (Carl M. Loeb University Professor Emeritus, Harvard Law School)
  • Lawrence Lessig (Roy L. Furman Professor of Law and Leadership, Harvard Law School)
  • Thomas Wolf (Co-founder and Chief Science Officer, Hugging Face)
  • Igor Babuschkin (Co-founder and CEO, River AI; Co-founder, xAI)
  • Rumman Chowdhury (Founder and CEO, Humane Intelligence PBC)
  • Tristan Harris (Co-founder and President, Center for Humane Technology)
  • Guillaume Verdon / "Beff Jezos" (Founder and CEO, Extropic; Founder of the e/acc movement)
  • Max Tegmark (Professor of Physics, MIT; Founder and Chair, Future of Life Institute)
  • George Hotz (Founder and CEO, the tiny corp; Founder, comma.ai)
  • Geoff Ralston (Founder, Safe AI Fund; Former President, Y Combinator)
  • Peter Scholze, James Maynard, and David Mumford (Fields Medalists in Mathematics)
  • Dan Hendrycks (Executive Director, Center for AI Safety)

“People in this coalition disagree on almost everything, yet we all recognize the danger of a company that convinces itself its own dominance is the greater good,” said Dan Hendrycks, Executive Director of the Center for AI Safety. “You cannot credibly claim to be safe while weakening your own previous safety commitments and asking the public for tens of billions of dollars to build recursively self-improving AI.”

Key Evidence: 

  • Racing Toward Recursive Self-Improvement: Anthropic explicitly warns that recursively self-improving AI may concentrate power and escape human control, while simultaneously pitching recursive self-improvement to public-market investors as one of its core competitive advantages.
  • Using Safety Branding to Court Investors, Employees, and the Public: Anthropic markets itself as an “AI safety and research company” to recruit safety-conscious talent and pitch investors as it attempts to automate AI R&D. 
  • Weakened Safety Commitments and Self-Serving Lobbying: In February, Anthropic weakened its Responsible Scaling Policy, including fully abandoning its unilateral commitments to pause. It has lobbied to weaken SB 1047, the RAISE Act, the AEGIS Act, and privately advocated for federal pre-emption of state AI laws.

The full statement and complete list of signatories can be viewed at misanthropicstatement.com.

About the Center for AI Safety: The Center for AI Safety (CAIS) is a San Francisco-based research and advocacy nonprofit dedicated to reducing societal-scale risks from advanced artificial intelligence. Through technical research, policy development, and public engagement, CAIS works to ensure AI is developed safely and remains under human control. Learn more at safe.ai.

Media Contact: media@safe.ai