A bipartisan group in the House has introduced legislation that would let the federal government shut down or restrict artificial intelligence systems judged to pose serious risks. The proposal arrives as concern over AI safety intensifies, with the reported OpenAI hack adding new urgency to the debate over how powerful models should be overseen.
The measure is aimed at the biggest players in the AI industry rather than smaller developers. Based on the details available, it would apply to companies generating at least $500 million a year from AI technology and would generally cover models built using at least $100 million in computing resources. Those thresholds suggest lawmakers are focusing on advanced, costly systems with the potential for broad impact.
Supporters appear to be framing the bill as an emergency tool for dealing with dangerous AI capabilities before harm spreads. By authorizing the government to throttle or halt certain models, the proposal would go beyond broad voluntary guidelines and move toward direct federal intervention when officials believe the risk is high enough.
The legislation also highlights how AI policy is becoming both a national security and technology governance issue on Capitol Hill. As lawmakers weigh the benefits of rapid innovation against fears of misuse, cyber vulnerabilities and loss of control, proposals like this one signal a tougher approach to regulating the most powerful AI systems.
