Anthropic co-founder calls for halting AI development without human oversight

Anthropic co-founder Jack Clark told the BBC that as AI models grow more capable, the risk of humans losing control becomes more concrete. He argued that the industry needs to make "halting progress without human consent" an institutional norm in the coming years, rather than a niche policy debate.

Anthropic separately published a policy paper proposing a framework for coordinated industry pauses if defined safety thresholds are crossed. According to the BBC, the document combines voluntary corporate commitments with a call for government oversight, though the verification mechanism is still open to debate among experts and policymakers.

OpenAI and Google DeepMind had not issued formal responses at the time of publication. Critics noted that voluntary commitments do not bind Chinese and other rival labs, while supporters argued the framework could begin to repair public trust. The White House's coming meeting with AI leaders will test whether the proposals translate into policy.