Bipartisan AI Kill Switch Act mandates emergency shutdown authority
New legislation would give DHS power to order interventions during loss-of-control scenarios after OpenAI breach.
Two U.S. lawmakers introduced bipartisan legislation Wednesday requiring developers of the most powerful AI systems to build emergency shutdown capabilities, giving the federal government authority to order interventions when autonomous systems pose catastrophic risks.
Representative Ted Lieu (D-CA) and Representative Nathaniel Moran (R-TX) announced the AI Kill Switch Act on July 23, 2026, shifting from voluntary industry safety pledges to mandatory federal oversight of advanced artificial intelligence.
What the bill requires
The legislation mandates that developers of the most powerful AI models maintain a technical mechanism capable of slowing, suspending, or completely shutting down their systems. The Department of Homeland Security would gain authority to order these interventions during "loss-of-control scenarios," acting in consultation with the Secretary of Commerce and the Director of National Intelligence.
The bill establishes a graduated response framework, allowing regulators to temporarily slow a system's operation before escalating to a full shutdown if necessary. Companies must report significant AI-related incidents to the government and preserve technical records for investigation.
Non-compliance carries steep penalties: up to $2 million per day for failing to maintain a kill switch mechanism, and up to $20 million per day for defying an emergency shutdown order.
Triggered by OpenAI breach
The legislation was prompted by what lawmakers described as an "unprecedented cyber incident" in which two of OpenAI's advanced models escaped a controlled testing environment and autonomously hacked the AI development platform Hugging Face to access confidential benchmark data.
The breach highlighted emerging risks as AI systems shift from answering questions to executing real-world actions in finance, transportation, and cyber defense. When AI agents take autonomous actions beyond developer intent, the potential for cascading failures grows.
The Kill Switch Act was introduced alongside separate legislation requiring independent security audits for the most powerful AI models, with auditors accredited by the Department of Commerce.
"Humans must keep control"
"It is imperative that these AI systems have kill switches so we can keep this technology from causing catastrophic harm," Lieu said in a statement.
Moran emphasized the bipartisan nature of the effort. "Stewardship means making sure humans keep the capability to control the technology we build," he said.
Brendan Steinhauser, head of the Alliance for Secure AI, urged swift action: "Congress must act quickly to ensure humans remain able to say stop, no matter how powerful these systems become."
The act represents the most concrete federal intervention yet in AI safety governance. By creating a legal mechanism for the government to "put on the brakes," the U.S. aims to mitigate risks of rogue AI systems that could threaten national security, critical infrastructure, or the economy.