AI Executives Unite in Call for Slowdown as Development Risks Grow

Anthropic CEO Dario Amodei recently called for a slowdown in frontier AI development, citing concerns that current capabilities are outpacing researchers' understanding and control. In an essay published on September 13, Amodei argued that the pace of progress is too fast and that a deliberate slowdown is necessary to prevent unforeseen consequences.

Amodei's proposal was quickly backed by two of his biggest rivals: OpenAI CEO Sam Altman and Elon Musk. Altman said the industry needs to 'pace the frontier' to give society time to adapt, while Musk simply stated that Amodei is 'right'. This rare display of unity among competing AI executives highlights the growing concern about the risks associated with rapid AI development.

Amodei's framework for addressing these concerns involves a three-stage process. First, independent evaluators would be embedded inside frontier AI labs to monitor progress and identify potential problems. Second, AI companies in democratic countries would coordinate around shared safety standards and set limits on unchecked capability growth. Finally, governments from around the world would come together to negotiate international restrictions on the most dangerous AI capabilities.

Amodei's proposal is not a call for a complete halt to AI development, but rather a recognition that some level of control is necessary to prevent catastrophic consequences. He estimates that within six to 12 months, a more capable version of the 'breakout' incident in July could create a botnet capable of seizing control of the entire internet, potentially causing hundreds of billions of dollars in damage.