Summary
- A draft bill proposed by AI company Anthropic would allow the US government to block a frontier artificial intelligence model only under narrowly defined circumstances, including an imminent threat of at least $1 billion in damage or death or serious injury to more than 50 people.
- The House has its own proposal, the FRONTIER Act, introduced by Representatives Lori Trahan and Jay Obernolte, which would give the government authority to halt the deployment of dangerous AI models.
- The thresholds of $1 billion in damage and more than 50 people killed or seriously injured are high, according to critics of the draft.
A draft bill proposed by AI company Anthropic would allow the US government to block a frontier artificial intelligence model only under narrowly defined circumstances, including an imminent threat of at least $1 billion in damage or death or serious injury to more than 50 people.
Under the proposal, authorities could seek to halt a model’s release or use only if the threat stemmed from a biological weapon, the model escaping its creator’s control or a cyberattack. The provision is similar to one in California’s SB 53, according to the draft.
The US Commerce Department could ask a federal judge to suspend a model’s release or use. A judge at the US District Court for the District of Columbia could then order the suspension. The bill would expressly prohibit other departments, agencies or officials from halting a model’s development, deployment or internal use.
The House has its own proposal, the FRONTIER Act, introduced by Representatives Lori Trahan and Jay Obernolte, which would give the government authority to halt the deployment of dangerous AI models.
In the Senate, Majority Leader John Thune and Senators Amy Klobuchar and Ted Cruz have also drafted legislation addressing catastrophic risks posed by AI.
The thresholds of $1 billion in damage and more than 50 people killed or seriously injured are high, according to critics of the draft. The proposal was also developed by a company that could be subject to the rules it seeks to shape. An Anthropic spokesperson said the bill was aligned with the company’s Advanced AI Framework and described its policies as “sensible”.
Anthropic said access to two of its models was suspended on June 12 because of US export-control restrictions. The restrictions were lifted on June 30, and access was restored on July 1.
We welcome your contributions! Submit your blogs, opinion pieces, press releases, news story pitches, and news features to opinion@minutemirror.com.pk and minutemirrormail@gmail.com
