# US proposes AI Kill Switch Act for dangerous systems Trump administration could gain authority to shut down risky AI. - **Source:** [Ars Technica - All content](https://brifing.si/source/b728770c-8dda-457a-8986-84879c9fc42c) - **Published:** 2026-07-23 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/e03fbaec-224d-47f4-87e7-6da6d987e4d1 ## AI Kill Switch Act proposed A new bill in the US would grant government officials authority to shut down AI systems 'capable of causing catastrophic harm.' If passed by Congress, both the Trump administration and future administrations could exercise this power. AI companies would be required to implement technical capabilities allowing system shutdowns or restrictions when ordered. ## Compliance requirements and penalties Under the proposed law, AI developers refusing shutdown orders could face fines up to $20 million per violation day. The authority would be given to the Department of Homeland Security through amendments to the 2002 Homeland Security Act. Companies must build technical infrastructure enabling remote system throttling or shutdown capabilities. ## Scope and limitations The legislation specifically targets AI systems with 'catastrophic harm potential.' The government could demand complete shutdowns, feature disabling, or access restrictions. Clear criteria for determining which systems qualify remain undefined, raising questions about implementation standards and potential overreach. ## Industry reactions anticipated The tech industry's response to this legislation remains uncertain. Some experts argue such regulations might stifle innovation, while others consider them necessary safety measures. Increased lobbying efforts from major tech companies are expected as the bill moves through legislative processes. ## Global implications This US initiative may inspire similar AI regulations worldwide. Major economies like China and the EU are likely to develop their own AI safety protocols. The move highlights growing need for international cooperation in establishing global AI safety standards and governance frameworks. ## Future development scenarios The legislation could fundamentally transform AI development practices. Companies would need to design systems with integrated 'kill switch' capabilities from inception. Ethical and safety-focused AI development approaches may gain prominence, with increased emphasis on auditability and control mechanisms in system architectures. ## Sources covering this story - https://arstechnica.com/tech-policy/2026/07/ai-kill-switch-act-would-let-trump-admin-order-shutdown-of-rogue-ai-systems/ — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._