Anthropic says AI labs need coordinated plan to halt development if risks rise - Reuters

Anthropic, the artificial intelligence company, called on frontier AI developers to establish a coordinated mechanism to slow or pause development if advanced systems begin showing signs of rapid self-improvement that could outpace societal risk management. The proposal was reported by Reuters on June 4, 2026.

According to Anthropic, such a plan should involve major AI labs agreeing on thresholds that trigger a slowdown or halt. The company emphasized that the pace of capability advances could lead to scenarios where society cannot adequately prepare for potential harms. Anthropic did not specify which thresholds or metrics would be used.

The proposal adds to ongoing discussions about AI governance, as regulators and industry groups debate the need for preemptive safety measures. Anthropic's call comes amid increasing attention to the risks of uncontrolled AI development, though no concrete agreement among labs has been announced. The company itself is a participant in the frontier AI race.