Anthropic Faces 89 Signatories Challenging Its AI Safety Claims Ahead of IPO

The MisAnthropic Statement alleges that Anthropic’s commercial ambitions and lobbying efforts conflict with its public safeguards, though the company had not responded as of 19 November 2026

Anthropic
89 signatories challenge Anthropic's AI safety claims, questioning its policies and lobbying ahead of its IPO Screenshot from KNWA & FOX24 - Northwest Arkansas & River Valley News/Facebook

A coalition of scientists, legal scholars and technology leaders has challenged Anthropic's public commitment to artificial intelligence safety, accusing the company of pursuing technology that could concentrate power and escape human control.

The Center for AI Safety (CAIS) announced the MisAnthropic Statement in a San Francisco-datelined release on 18 November 2026.

To recall, Anthropic has made AI safety a central part of its public identity. The statement argues that this positioning conflicts with the company's efforts to develop recursively self-improving AI, which its signatories say could create risks even Anthropic itself has warned about.

A contemporaneous count put the coalition at 89 signatories, although the online list may change. Business Wire did not provide a definitive total.

Why Are 89 Signatories Challenging Anthropic?

The MisAnthropic Statement brings together AI safety researchers, open-source advocates, AI ethicists and accelerationists.

The group does not share a single position on how AI should be developed, but its members have joined in criticising Anthropic's safety branding and commercial ambitions.

The statement says Anthropic warns that recursively self-improving AI could concentrate power and escape human control, yet is racing to build it. It accuses the company of using its safety reputation to attract investors, employees and public support while weakening its own safeguards.

The statement also accuses Anthropic of lobbying for regulations that favour its interests. These remain allegations by the signatories, not established findings by regulators or courts.

The signatories include Harvard academics Steven Pinker, Oliver Hart, Judea Pearl, Laurence Tribe and Lawrence Lessig. Technology figures include Hugging Face's Thomas Wolf, Igor Babuschkin of River AI and xAI, MIT's Max Tegmark, Tristan Harris of the Center for Humane Technology and George Hotz of the tiny corp.

Dan Hendrycks, executive director of CAIS, said the coalition brought together people with sharply different views on AI policy.

'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,' he said.

Anthropic's Safety Policy and Lobbying Under Scrutiny

The dispute also concerns Anthropic's Responsible Scaling Policy, its framework for addressing risks from increasingly capable AI systems. CAIS alleges that Anthropic weakened the policy in February, including by abandoning previous unilateral commitments to pause development.

Anthropic introduced Version 3.0 of its Responsible Scaling Policy on 24 February 2026, adding Frontier Safety Roadmaps and risk reports. On 2 April, Version 3.1 clarified that the company could still pause AI development at its discretion, even when the policy did not require it.

This provides important context to CAIS's criticism, but does not resolve the dispute over how strongly Anthropic is committed to pausing development.

CAIS also accuses Anthropic of lobbying against California's SB 1047, the RAISE Act and the AEGIS Act. It further alleges that the company privately pushed for federal pre-emption of state AI laws.

The criticism comes ahead of Anthropic's planned IPO, with CAIS arguing that the company uses its safety reputation to attract investors while pursuing recursively self-improving AI.

As of 19 November 2026, the Business Wire announcement contains no response from Anthropic to the allegations. The statement and its evolving list of signatories remain available online.