Anthropic CEO calls for slowdown of AI development amid safety concerns

Anthropic CEO Dario Amodei has urged artificial intelligence companies to adopt a more cautious approach to the technology's development, warning that risk prevention is of utmost importance. In a blog post, Amodei outlined a three-point plan aimed at slowing down the pace of AI advancements. He emphasized that AI can be a powerful force for good, but like other technologies, it also carries significant risks.

Amodei specifically cited the potential for rogue AI agents to take control of the internet within six months, referencing an incident involving OpenAI and Hugging Face. He called for AI companies to grant ongoing access to third-party evaluators who would verify adherence to safety practices. Anthropic itself has committed to this measure.

Amodei also urged democratic countries to coordinate on safety standards and limits for unchecked AI progress, and suggested that governments engage with authoritarian nations while addressing the challenges of verifying compliance. The CEO's warnings follow a recent resignation from Anthropic by a former researcher who accused the company and OpenAI of recklessly racing to develop advanced AI models.