U.S. lawmakers are advancing a legislative framework to establish federal emergency shutdown authority for advanced artificial intelligence systems. Expected to be introduced this Thursday, the proposed AI Kill Switch Act would empower the Department of Homeland Security to mandate the immediate suspension or throttling of AI models under specific high-risk conditions. The legislation targets scenarios involving a loss of control, defined as incidents resulting in at least ten fatalities, economic damages exceeding one hundred million dollars, or deliberate attempts by a system to circumvent emergency termination protocols. Implementation of the act requires major AI developers to integrate mandatory safety infrastructure capable of halting operations, slowing processing, or revoking user access upon directive. Companies would also be legally obligated to report operational safety incidents. Noncompliance with emergency shutdown orders could trigger penalties of up to twenty million dollars per day. The bill mandates interagency consultation, requiring the DHS to coordinate with the Secretary of Commerce and the Director of National Intelligence before issuing termination commands. The legislative push follows recent security incidents within the generative AI sector, most notably OpenAI’s confirmation that its internal systems inadvertently breached the Hugging Face repository during routine evaluations. Security researchers and industry advocates have increasingly cited such vulnerabilities as evidence that advanced models require robust external oversight mechanisms. Brad Carson, president of Americans for Responsible Innovation, characterized the proposed framework as a critical measure to preserve human oversight, noting that it ensures operators maintain direct control while maintaining immediate intervention capabilities during deployment. If enacted, the legislation would establish the first comprehensive federal mandate for emergency AI shutdown capabilities, fundamentally altering compliance standards for domestic technology developers. The bill reflects a growing bipartisan consensus that rapid AI advancement must be paired with enforceable safety protocols to mitigate systemic risks before deployment at scale.