Anthropic CEO calls for FAA-style regulation of frontier AI releases

Anthropic co-founder and CEO Dario Amodei called for government regulation of powerful AI model releases in a new essay titled "Policy on the AI Exponential," comparing the industry to commercial aviation overseen by the Federal Aviation Administration. Frontier models, he argued, should go through technical testing and auditing, with releases blocked or reversed if they fail safety standards. Alongside the essay, Anthropic published an Advanced AI Framework targeting catastrophic model risks and an Economic Policy Framework on AI-driven labor displacement, backed by $350 million in new funding. The company proposes mandatory third-party testing for models trained using more than 10^25 floating-point operations, or developed by companies with over $500 million in AI revenue or $1 billion in AI R&D. The essay landed a day after Anthropic released Claude Fable 5, which the company describes as its most powerful general release model.