Politics and regulation·October 20, 2026, 02:47
US bans Chinese AI from all civilian government IT systems
AI-generated from the sources listed, with automated source checks.
The US Department of Homeland Security has ordered DeepSeek's AI removed from civilian federal agencies. The ban also covers products from other companies built on top of DeepSeek's models.

The US Department of Homeland Security (DHS) has ordered that technology from the Chinese AI provider DeepSeek be removed from all civilian federal agencies, PYMNTS reports.
The directive, titled Exclusion and Removal Order #2026-01, was signed by Homeland Security Secretary Markwayne Mullin. It was presented on October 19, 2026, and publication in the Federal Register, the official US government gazette, was scheduled for October 20. According to Gizmodo, Mullin's accompanying note is dated July, so the order may have been signed earlier than it was published.
What must agencies do?
Agencies must stop buying DeepSeek products and services and remove the technology they already have in their IT systems. The order is based on a 2018 US law on federal supply chain security.
The ban reaches further than DeepSeek itself. According to Security Magazine, it also covers third-party products and services developed on the basis of DeepSeek's models or implementations. That means software from another provider can also be affected if it builds on DeepSeek.
Gizmodo reports that the ban includes the State Department, among others. The Navy and the Department of Commerce already had their own restrictions.
DHS' reasons
DHS writes that DeepSeek's models pose "multiple threat vectors," meaning several possible avenues of attack. Among other things, under Chinese law the Chinese government can gain access to private companies' infrastructure.
DHS also points out that personally identifiable information is collected when an account is created, and that user data is sent to the model when users ask questions. DHS also cites security vulnerabilities in DeepSeek's iOS app. According to DHS's assessment, information transferred to China could provide usable intelligence on American individuals and government programs.
According to Gizmodo, Mullin says that account creation collects personally identifiable information and can also track and digitally fingerprint users. Gizmodo also mentions fears of malware in the models themselves.
Doubts about enforcement
Gabrielle Hempel of the security firm Exabeam calls the breadth of the order, which extends into the software supply chain, particularly interesting. According to her, it is unclear how the requirements can be enforced technically.
Deadlines for removal and any exemptions are not stated in the sources. Neither is DeepSeek's response.
Not the first ban
The Department of Commerce introduced a similar ban in March 2025. That same year, Australia, Taiwan and Italy banned DeepSeek, and Texas was the first US state to ban it on government devices.
Sources
Found via SiliconANGLE
Get the week's AI news in your inbox
Choose your level, topics and length. One email a week, unsubscribe at any time.
Subscribe to Promptly Newsletter



