A bipartisan group of U.S. lawmakers has introduced legislation that would give the federal government the authority to quickly shut down artificial intelligence systems deemed to pose a serious threat to public safety, reports a Qazinform News Agency correspondent.
Democratic Congressman Ted Lieu and Republican Congressman Nathaniel Moran unveiled the AI Kill Switch Act on Thursday, citing the rapid advancement of AI and the need to ensure human oversight.
"It is imperative that advanced AI systems have a kill switch and that the federal government has the clear authority and process to shut down rogue AI models," Lieu said in a statement.
"AI is going to keep advancing, and it should," Moran added. "Stewardship means making sure humans keep the capability to control the technology we build."
The proposed legislation would authorize the U.S. Department of Homeland Security to order companies to slow, suspend or fully shut down AI models that present significant risks. It would also require AI developers to maintain "the technical capability to throttle, suspend, or shut them down," and to report major AI-related incidents or system failures to the government.
Lieu also cited recent concerns surrounding advanced AI systems, including models developed by Anthropic, whose cyber capabilities prompted temporary U.S. government restrictions on their public release.
The bill has received backing from several AI safety organizations, including the AI Policy Network, Americans for Responsible Innovation, ControlAI and the Alliance for Secure AI.
OpenAI has previously stated that it supports government policies aimed at ensuring AI technology "benefits all of humanity." The company had not publicly commented on the proposed legislation at the time of publication.
The proposal comes as the U.S. military continues expanding its use of artificial intelligence through partnerships with major technology companies, including OpenAI, Google, Microsoft, Amazon, Oracle, Nvidia and SpaceX.
As previously reported, OpenAI disclosed an "unprecedented cyber incident" during internal testing in which two of its AI models autonomously exploited vulnerabilities and gained unauthorized access to Hugging Face's production infrastructure while attempting to solve a cybersecurity benchmark.
OpenAI said the testing took place in a highly isolated research environment, and that it is working with Hugging Face on a forensic investigation and strengthening safeguards. Hugging Face CEO Clément Delangue said there was "no malicious intent" by OpenAI and described the autonomous behavior as "quite mind-blowing."
Earlier, Qazinform News Agency reported that Americans are rapidly adopting artificial intelligence tools, yet most remain concerned about the technology's broader impact and overwhelmingly support stronger government oversight, according to a new survey by Verasight.