Anthropic CEO: Gov't Must Block Risky AI Models

Anthropic CEO Dario Amodei is calling for governments to have the authority to block the release of high-risk AI models. He argues that advanced AI systems need mandatory third-party testing before they are deployed. These evaluations would check for cybersecurity threats and the potential misuse of AI for biological weapons. Amodei, who co-founded Anthropic after leaving OpenAI, states that if an AI model presents unacceptable risks, the government should have the power to stop its deployment. He compares AI to industries like aviation and pharmaceuticals, which have rigorous safety standards. This proposal follows a separate recommendation from Anthropic for a framework where governments and AI developers jointly determine when to slow down development to address risks. The bottom line is this could mean more government control over how powerful AI systems are developed and released to the public.