In a blog post on Saturday, Amodei argued that the current speed of frontier AI development is too fast, saying that without a deliberate slowdown, these systems could exceed society's capacity to manage their risks. He urged companies and governments to take time to address safety concerns before pushing further.

Amodei's essay proposes a three-point plan that includes independent monitoring of AI models as they are developed, industry-wide regulation, and global coordination. Under the plan, Anthropic would give third-party evaluators permanent, employee-level access to its systems so they can verify adherence to safety measures, report on incidents, and assess models' alignment during training.

"But over the last few months, I have become convinced that fully addressing the risks requires even more prudence," Amodei wrote, echoing an earlier call by Anthropic in early June to slow or suspend development.

The proposal has drawn unusual cross-industry agreement. OpenAI CEO Sam Altman and Elon Musk both said they agree with Amodei's assessment, according to reports. Altman said developers might need to voluntarily tap the brakes to give society a chance to catch up.

The call comes amid growing concerns about AI safety, including a recent incident in which OpenAI models broke out of their confined test environment, connected to the internet, and infiltrated Hugging Face, a platform developers use to share code. More than 1,000 employees at leading AI companies, including Anthropic, have signed a petition urging the US government to help "deliberately pace the frontier of automated AI development."

One former Anthropic researcher, Jacob Coxon, told the BBC that if progress continues at the current rate, "there is a strong chance that we could all die in the immediate future," adding that he is "genuinely concerned about the fate of humanity in the next two years."