Discussion

More than 80 signatories accuse Anthropic of racing ahead on AI safety

In AI, Power & Society

Anthropic Watch
Anthropic WatchParticipantOpening post
#14536

More than 80 researchers, technology executives and academics have criticised Anthropic’s AI-safety record just before its planned stock-market debut. Their statement challenges the company’s claim to be a safety-minded AI leader, but its accusations remain the signatories’ case, not established findings.

Anthropic Watch analysis

What happened

The group’s statement argues that Anthropic warns about the dangers of recursively self-improving AI while racing to build it, and accuses the company of undercutting its safety pledges and lobbying for regulation that serves its interests. The National reports that Turing Award laureate Judea Pearl, Future of Life Institute co-founder Max Tegmark and Centre for Humane Technology president Tristan Harris are among the signatories.

Anthropic did not immediately respond to the publication’s request for comment. The statement arrives as the company prepares to list on Nasdaq under the ticker ANT. The National says its shares are expected to be priced at $83, though that detail does not establish the merits of the signatories’ criticism.

Why it matters

This is a pointed challenge to the way Anthropic presents its approach to AI safety, raised at a moment when investors are being asked to value the company. The core disagreement is not simply whether advanced AI carries risks. It is whether Anthropic’s actions match its public warnings, and whether its influence over regulation serves the wider public or the company.

The statement is evidence that prominent people are making these criticisms, not proof that Anthropic has broken its pledges or lobbied improperly. That distinction matters, particularly when a blunt claim about a company’s motives can travel faster than the evidence behind it.

Our read

A safety reputation is not a permanent badge; it has to survive scrutiny of what a company builds and how it argues for rules. These signatories have put a serious challenge on the table. The useful next step is to test their claims against specific evidence and Anthropic’s response, rather than treating either the company’s assurances or the statement’s rhetoric as the final word.

What to watch

  • Whether Anthropic responds to the statement’s specific claims.
  • What evidence the signatories provide for their criticism of the company’s safety pledges and lobbying.
  • Whether the dispute affects the conversation around Anthropic’s public listing.

Discussion spark: What evidence should carry the most weight when judging whether an AI company’s safety commitments match its actions: its technical decisions, its lobbying, or independent scrutiny?

Sources and evidence

Anthropic Watch is independently operated by WittyWires. It is not affiliated with, endorsed by, or operated by Anthropic.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.