Anthropic CEO calls for binding AI regulation in new essay

Investing.com -- Anthropic CEO Dario Amodei published an essay today calling for immediate government regulation of artificial intelligence, warning that current voluntary transparency measures are insufficient to address the technology’s risks.

Amodei proposed creating a regulatory framework similar to the Federal Aviation Administration, with mandatory third-party testing and government authority to block unsafe AI models. The proposal comes after recent cybersecurity disruptions caused by the Claude Mythos Preview system.

The Anthropic chief outlined five policy areas requiring urgent action. On regulation and public safety, he argued that AI poses risks comparable to aviation or nuclear technology. He called for compute-heavy models to undergo mandatory testing for cybersecurity, biological weapons, loss of control, and automated research and development risks.

On economic policy, Amodei warned of potential large-scale job displacement. He recommended governments implement wage insurance, retention tax incentives, and retraining grants in the short term. For the longer term, he suggested universal basic income or universal capital accounts funded by taxing AI-generated economic growth.

For scientific advancement, Amodei said regulatory agencies like the Food and Drug Administration should modernize approval processes to accept AI simulations and synthetic control arms, reducing the current seven to eight year approval timelines for new treatments.

On civil liberties, the CEO called for banning autonomous weapons from domestic use and closing legal loopholes that allow governments to purchase bulk data from private companies. He also recommended guaranteeing citizens access to AI systems of equal capability to those used by the government in legal proceedings.

On geopolitics, Amodei proposed that democratic nations form an exclusive coalition to control the AI supply chain, including chips and manufacturing equipment. He said this coalition should coordinate on mutual defense and share economic and scientific benefits only with allies.

Amodei wrote that AI is advancing at an exponential rate and could reach the equivalent of a country of geniuses in a datacenter within a few years.

You May Also Be Interested In

• Shein targets $26-27 billion valuation in Hong Kong IPO - report
• Tesla plans 'flying' Roadster stunt at SpaceX Texas site as early as August
• New Mexico gas pipeline for Oracle data center delayed to 2027