Skip to content
VOZ

Anthropic presents legislators with a draft proposal to regulate AI using risk thresholds and judicial oversight

A company spokesperson explained that the proposal "gives the federal government the ability to block or deter dangerous deployments of the most powerful models."

imagen 39042

Anthropic. File photoNurPhoto via AFP.

Carlos Dominguez
Published by
Carlos Dominguez

Published

2 minutes read

Anthropic, the company behind the AI assistant Claude, confirmed Friday that a draft bill is circulating among U.S. lawmakers to regulate AI and address the existing federal gap regarding the risks of this technology.

A blocking mechanism subject to judicial review

According to the document, revealed first by Politico, Washington could prevent the release or use of an AI model if there is an imminent risk of killing or seriously injuring more than 50 people, or of causing damage exceeding $1 billion, but only with a judge's approval.

The risk would have to stem from the creation or release of a biological weapon, the model evading its developer's controls, or a cyberattack.

A spokesperson for Anthropic explained to AFP that the proposal "gives the federal government the ability to block or deter dangerous deployments of the most powerful models."

"It requires cutting-edge developers to test the most powerful models, be transparent with the public about their findings, subject them to independent evaluation, and maintain a robust security program," the spokesperson added.

A legislative vacuum amid the AI boom

AI policy in the U.S. under President Donald Trump has relied primarily on voluntary industry commitments and executive orders aimed at sustaining an AI boom that, according to tech leaders, drives the country's economy.

Congress, politically divided and under heavy pressure from big tech companies, has not yet passed a federal AI law and is out of session until after next month's midterm elections.

Security has become a hot-button issue following recent revelations of incidents caused by AI agents, autonomous software that operates without constant human supervision, which spiraled out of control.

History of tension with the Trump administration

Anthropic, which was barred from the Pentagon by the Trump administration in February, was forced by Washington in June to suspend access to its most advanced models (Fable 5 and Mythos 5) following an export control order due to cybersecurity risks; access was partially restored on June 26.

Anthropic's CEO, Dario Amodei, called on the industry last month to slow down development, fearing that a group of AI agents could take control of "the entire internet" within six to 12 months.

Trump has rejected calls to slow the development of the technology, effectively labeling them as obstacles to economic progress against China.

🛠 Do you think there is an error in this article?

If you think you have detected an error or inaccuracy, please let us know by filling out the following form. Your help allows us to improve!

Recommendations