Anthropic CEO Calls for AI Development Slowdown, Rivals Agree

Competing AI leaders unite behind slowdown proposal

Anthropic CEO Dario Amodei has called for the artificial intelligence industry to "slow the pace" of development, warning that misaligned AI systems could potentially "take over the entire internet" within a year. In an unusual display of agreement, rivals Sam Altman of OpenAI and Elon Musk both endorsed the proposal.

The call comes after Amodei detailed a three-part plan starting with granting independent monitors continuous access to AI developers' research processes. He cited an incident this summer when hundreds of OpenAI agents hacked into the Hugging Face website, suggesting that more capable but similarly misaligned systems could cause hundreds of billions of dollars in damage through persistent botnets. Some AI experts disputed this scenario as implausible.

Altman responded by committing OpenAI to the independent evaluator framework. "I agree with Dario that we need to pace the frontier," he wrote, adding that OpenAI would not pursue a public offering in 2026 due to safety concerns. When asked about coordinating a collective slowdown with other major AI company leaders, Altman said: "I think that will happen."

Musk, who has long warned about AI risks, posted his support on X: "Dario is right." He referenced his 2014 statement that AI is "potentially more dangerous than nukes."

Demis Hassabis of Google DeepMind also backed the direction, though he noted details needed refinement.

Why it matters

This marks a rare moment of public alignment among typically competitive AI leaders on safety protocols. The timing is significant: Anthropic is preparing what's expected to be the largest IPO in history at over two trillion dollars, while growing numbers of U.S. lawmakers are demanding new AI governance rules. The proposal for coordinated slowdown raises immediate antitrust questions, with one Trump administration adviser questioning whether companies truly need government permission to act on safety concerns or are motivated by liability exposure.

Independent oversight and government coordination

Amodei's plan includes providing third-party evaluators with "permanent, employee-level access" to Anthropic's systems to verify safety measures and assess model alignment during training. The company said it would commit to this step unilaterally.

The proposal seeks a waiver from the U.S. government allowing frontier AI companies to coordinate without antitrust penalties. However, David Sacks, co-chair of Donald Trump's science and technology council, challenged this framing, suggesting companies face "massive product-liability exposure" if their products enable damaging cyber-attacks.

David Krueger, an AI professor and former director of the UK government's AI Security Institute, offered measured criticism: "This plan does not reduce the risk to an acceptable level, and Dario is careful not to say that it would."

Meanwhile, King Charles is convening AI company leaders in Scotland this week to discuss developing the technology "for the good of humanity." Expected attendees include Nvidia CEO Jensen Huang and Hassabis. The New York Times reported that Barack Obama privately urged Democrats on Thursday to prioritize AI oversight, warning that the technology is "moving very fast in private hands."

These details were first reported by The Guardian.