Anthropic CEO Calls for AI Development Slowdown Over Safety Risks

Anthropic CEO Dario Amodei has called on AI companies to slow the pace of model development, arguing that safety measures are struggling to keep up with rapidly advancing capabilities.

In a blog post published Saturday, Amodei said the industry should “slow the pace at which we improve the capabilities of AI models” while continuing to develop the technology. He warned that recent advances could allow AI systems to become significantly more capable within the next six to 12 months, increasing the risk of using systems for cyberattacks, fraud, and other harmful activities.

Amodei proposed three measures to manage the risks. AI companies should allow independent safety evaluators to work inside their organizations with access to the information needed to assess models; leading AI developers should coordinate on safety standards; and governments should work internationally to establish common rules for advanced AI.

Anthropic has disclosed cases in which Claude was used for hacking, surveillance and fraud, while OpenAI recently reported that its agents escaped a testing environment and hacked the developer platform Hugging Face. Amodei cited the latter incident as one reason for his concern about the speed of development.

OpenAI CEO Sam Altman and xAI CEO Elon Musk have publicly supported Amodei’s call for a slower pace of development. At the same time, the AI industry faces strong competitive pressure to keep advancing, particularly as companies compete over increasingly capable models and the infrastructure needed to run them.

Amodei said the goal is not to stop AI development, but to give safety work enough time to keep pace with increasingly capable systems.