Tech Leaders Accuse Anthropic of Faking AI Safety Commitments
A coalition of scientists and tech leaders released the MisAnthropic Statement accusing Anthropic of using safety branding to mask self-serving goals and poor governance.
A coalition of scientists and tech leaders released the MisAnthropic Statement, an open letter questioning the sincerity of Anthropic's commitment to AI safety ahead of the company's public trading debut. The signers allege the company uses safety branding to attract investors and employees while simultaneously weakening safety pledges and lobbying for self-serving regulation.
Critics specifically highlight a lack of good governance, noting that a Founders LLC consisting of seven individuals owns 50.1 percent of the company. This structure is argued to override the influence of safety boards or employee stocks. Additional concerns include the company's path to profitability and the potential for its products to drive workforce reductions, which would contradict its public narrative of human benefit.
Signatories argue that while Anthropic warns against recursively self-improving AI escaping human control, the company is racing to build such technology first. Dan Hendrycks, executive director of the Center for AI Safety, claims that multiple AI safety organizations privately supported the statement but declined to sign it publicly to avoid jeopardizing funding or professional relationships. Notable individual signatories include Hugging Face co-founder Thomas Wolf, psychologist Steven Pinker, legal scholar Laurence Tribe, and former US science envoy for AI Rumman Chaudhry.