Anthropic urges government power to block dangerous AI

Anthropic has released two new policy blueprints that call on governments to assume direct legal authority to block dangerous AI deployments and to establish economic protections for workers as artificial intelligence reshapes employment. The company’s Advanced AI Framework focuses squarely on safety regulation, while the companion Economic Policy Framework tackles workforce displacement, capital distribution, and the strength of the social safety net. Together, the documents represent one of the most detailed proposals yet from a leading AI developer for how public institutions should oversee the technology. The safety framework argues that current regulatory tools are insufficient and that governments must be empowered to halt releases that pose unacceptable risks, even after models have been built and tested. On the economic side, Anthropic warns that without proactive policy interventions, the benefits of AI will concentrate among a small group of owners and shareholders, leaving millions of workers without viable income pathways. The proposals arrive as lawmakers in multiple countries debate how to balance innovation with public protection, and they signal that some industry players are now pushing for stronger, more explicit government authority rather than relying solely on voluntary commitments.