AI News

Anthropic Faces ‘Not a Safe AI Company’ Accusation From 89 Signatories Ahead of Nasdaq Debut 

Anthropic Faces ‘Not a Safe AI Company’ Accusation From 89 Signatories Ahead of Nasdaq Debut
Anthropic Faces ‘Not a Safe AI Company’ Accusation From 89 Signatories Ahead of Nasdaq Debut

Key Highlights:

  • A coalition of AI researchers, Nobel laureates, technology founders and legal scholars has accused the AI company of using its AI safety reputation to advance commercial and political interests.
  • The November 18 statement, called the MisAnthropic Statement, initially attracted 89 signatories, including Steven Pinker, Max Tegmark, Igor Babuschkin, Thomas Wolf, George Hotz and Gary Marcus. The list had grown to 92 by early November 19.
  • The signatories allege that Anthropic is pursuing recursively self-improving AI despite warning about its potential risks, while weakening safety commitments and lobbying for regulations that serve its interests.

Anthropic is facing criticism from a coalition of prominent researchers, technology founders and academics just as the artificial intelligence company prepares to make its Nasdaq debut. The group has published a statement declaring that Anthropic is “not a safe AI company,” challenging the organisation’s public positioning as a leader in responsible AI development. Published on November 18, the MisAnthropic Statement brings together 89 signatories at the time of its initial circulation, with the number increasing to 92 by early November 19. 

The criticism comes at a significant moment for Anthropic, which priced its initial public offering at $83 per share on November 18, ahead of its expected first day of trading under the ticker ANT on November 19. The signatories argue that the company’s public safety messaging conflicts with its commercial ambitions and its pursuit of increasingly capable AI systems. 

Why Are Researchers Challenging Anthropic’s AI Safety Claims?

The coalition’s statement centres on what it describes as a contradiction between Anthropic’s warnings about advanced AI risks and its own efforts to develop more powerful systems.

The statement says: “Anthropic warns that recursively self-improving AI could concentrate power and escape human control, yet races to build it first.” It further accuses the company of using “safety branding to court investors, employees, and the public,” while allegedly weakening its previous safety commitments and lobbying for regulation that benefits its commercial interests. 

It concludes by declaring that Anthropic is not a safe AI company and calls for the rejection of what the signatories describe as its attempt to accumulate power in the name of the greater good. The statement has attracted support from individuals across different areas of the AI industry, academia and technology policy. Signatories include Steven Pinker, Harvard professor of psychology; Max Tegmark, professor of physics at MIT and founder of the Future of Life Institute; Thomas Wolf, co-founder and chief science officer of Hugging Face; and Igor Babuschkin, co-founder of xAI and CEO of River AI.

Other prominent names include George Hotz, founder of comma.ai and the tiny corp; Gary Marcus, an AI researcher and author; Guillaume Verdon, also known as Beff Jezos and founder of Extropic; and Tristan Harris, co-founder of the Center for Humane Technology. The coalition also includes Nobel laureate Oliver Hart, Turing Award laureate Judea Pearl, Harvard legal scholars Laurence Tribe and Lawrence Lessig, and Fields Medalists Peter Scholze, James Maynard and David Mumford. Dan Hendrycks, executive director of the Center for AI Safety (CAIS), is also among the signatories. 

The range of signatories is notable because the coalition includes people with differing views on AI development. It brings together AI safety advocates, open-source supporters, researchers and figures associated with effective accelerationism. According to the Center for AI Safety, the participants share opposition to what they view as Anthropic’s attempt to use its safety reputation to strengthen its influence.

Hendrycks said the coalition members disagree on many issues but share concerns about a company that believes its own dominance serves the greater good. He also questioned whether Anthropic could credibly present itself as a safety-focused organisation while weakening earlier commitments and seeking substantial public-market funding to develop recursively self-improving AI. 

What Are the Allegations About Anthropic’s Safety Pledges and Lobbying?

Beyond the backlash of Anthropic’s AI development strategy, the statement raises concerns about changes to the company’s Responsible Scaling Policy ( RSP) and its involvement in AI regulation. The Center for AI Safety’s announcement identifies the February revision of Anthropic’s Responsible Scaling Policy as a core issue. As per the coalition, the company weakened its prior commitments including abandoning unilateral promises to pause development when it could not guarantee that advanced AI systems would remain safe.

The Future of Life Institute (FLI), whose co-founders Max Tegmark and Anthony Aguirre signed the statement, has also published a separate explanation of its concerns. It argues that recursive self-improvement could allow AI systems to develop more capable successors. It might potentially accelerate technological progress beyond human supervision.The institute further argues that Anthropic’s pursuit of automated AI development conflicts with its public calls for caution. Tegmark has criticised the company’s approach, saying its public warnings about uncontrolled AI development do not resonate with what he describes as its efforts to accelerate the same technology.

The allegations have not established that Anthropic violated any law or that its AI systems are unsafe. They represent the signatories’ criticism of the company’s policies, development strategy and public messaging. The statement also does not provide detailed evidence for every allegation about weakened pledges and lobbying. The debate nevertheless puts Anthropic’s safety commitments under scrutiny as the company begins its next phase as a publicly traded AI developer.

Khwaish Manwani
Khwaish Manwani, an inquisitive soul fond of words and driven by a profound interest in article writing that brings thoughts to life. Apart from her way with the words, she also pursues table tennis as a side passion.
You may also like
More in:AI News