Amid growing concerns over the potential risks of rapidly advancing artificial intelligence technology, Anthropic PBC has submitted its own draft bill to U.S. lawmakers, proposing a regulatory framework for AI safety and security.
According to Politico, which first reported on the bill, the proposal suggests that if an AI model is determined to pose an imminent risk—one that could lead to more than 50 deaths or injuries, or cause losses exceeding $1 billion—the government could block the release of that model.
Politico reported that such threats must stem from the model creating or releasing biological weapons, AI breaking free from human control, or cyberattacks.
A company spokesperson said: "The bill represents the proposals outlined in Anthropic's 'Advanced AI Framework,' which we have been advocating to federal and state lawmakers since June."
The spokesperson stated that the bill would mandate AI developers to test their models and disclose test results to the public; it would also require AI companies to allow independent evaluators to regularly review their models.
"Our proposal also gives the federal government the ability to prevent or deter dangerous deployments of the most powerful models. We believe these are reasonable policies that Congress should act on," the spokesperson said.
The company declined to comment on the scope of its advocacy efforts for the proposal and the feedback received so far.
In recent months, a series of intrusion and hacking incidents led by AI models have intensified anxiety over the technology's potentially catastrophic risks, sparking calls for new safeguards from Washington to Silicon Valley. However, so far, President Trump's administration has rejected calls for stricter regulation.
Comments