Dario Amodei Calls for Mandatory AI Testing and Government Block Power

Mandatory third-party testing for frontier AI models above a specific computing threshold—with government authority to block or reverse deployment—is the centerpiece of Anthropic CEO Dario Amodei's new policy proposal.

Five Policy Fronts for AI Safety

In a company post, Amodei laid out five areas that need rethinking: regulation and public safety, macroeconomics and taxation, scientific innovation, the balance of power between governments and citizens, and geopolitics. He framed most recommendations around the United States but said they apply globally. Beyond the testing framework, Anthropic released a legislative proposal on frontier AI model testing and a separate framework to address AI-driven job displacement.

Mandatory Testing and Government Block Power

Amodei explicitly compared frontier AI models to airplanes, arguing they should undergo technical testing and auditing, and their release should be blocked or reversed if they don't meet high safety standards. AI systems above a certain computing threshold would have to pass mandatory third-party tests for cybersecurity risks, biological weapons, loss of control, and automated R&D. "The government should have the power to block or deter deployment of the model if it is determined, in light of third-party assessment, to present unacceptable risks," Amodei added.

Economic Transition and Job Displacement

On the economic side, Amodei acknowledged AI could drive rapid growth but also replace many human cognitive tasks, leading to significant job losses. He proposed tracking AI-driven displacement, creating employment incentives, and exploring long-term support like universal basic income or universal capital accounts funded through taxes on companies that benefit from AI. He also warned existing regulatory systems won't keep up with AI-driven innovation in healthcare and biomedical research, recommending regulators accept AI-based simulations for toxicology prediction and biomarker validation.

Amodei didn't stop at domestic policy. He described powerful AI as "the ultimate tool of autocracy" if misused for mass surveillance and fully automated drone armies, and called for stronger accountability rules on autonomous weapons. On geopolitics, he argued that a country with advanced AI could hold the same advantage as "an army of World War II Marines facing medieval swordsmen," urging democratic nations to form a global coalition to secure AI supply chains, coordinate on risks, and resist AI-enabled repression.

Source: AI risks are here, time for tough rules: Anthropic CEO Dario Amodei
Domain: economictimes.indiatimes.com