Dario Amodei, Anthropic’s chief executive, has called for slowing AI development so that safety measures can keep pace with model capabilities. He warns that within 6 to 12 months of his September 12, 2026 essay, new models could enable swarms of AI agents to cause serious damage through cyberattacks, potentially amounting to hundreds of billions. His position has put him at odds privately with Jensen Huang, Nvidia’s chief executive, and Satya Nadella, Microsoft’s chief executive. At a White House event on September 29, 2026, at least 24 AI industry leaders signed voluntary safety principles, but Huang and Mark Zuckerberg defended rapid innovation, internal caution and market-based approaches rather than new laws. The disagreement could send conflicting messages to policymakers about whether AI requires stronger regulation.
Source: Read the original article


