Trump administration orders ban on DeepSeek in federal agencies
SiliconANGLE James Farrell ● Covered by 3 sources
The DHS ordered DeepSeek removed from federal civilian agencies. Washington’s treating the Chinese AI app like a security risk, not a tool.
Based on reporting by SiliconANGLE, James Farrell — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
The Trump administration has ordered DeepSeek out of federal civilian agencies. The Department of Homeland Security published the directive in the Federal Register on October 19, saying the Chinese company’s AI models pose “multiple threat vectors” and could give the Chinese Communist Party access to government infrastructure.
The ban goes beyond the app itself. Civilian agency systems will have to strip out all products tied to DeepSeek, including third-party services built with it. Future purchases of DeepSeek products and services are now off-limits in those agencies as well.
DHS leaned on the 2018 Federal Acquisition Supply Chain Security Act after a review by the Federal Acquisition Security Council. The agency said the technology could be used to “track or digitally fingerprint” users and might send sensitive data to the People’s Republic of China if used by government staff or vendors.
The crackdown has been building for a while. DeepSeek drew heavy attention last year when it released its R1 model, which was trained on less powerful chips and at a fraction of the cost of some competing systems, yet still beat models from OpenAI and Anthropic in several complex tests. That success did not calm U.S. officials down.
The Commerce Department banned DeepSeek from government devices last year, the U.S. Navy told its members not to use it, and New York State blocked it from government devices and networks over data privacy and censorship concerns. DeepSeek’s latest model, V4.1 Flash-1017, was released recently and was said to offer faster performance plus better reasoning and coding, with early reports suggesting it could rival some of the most advanced American systems.
My take — AI-written commentary, not fact-checked reporting
DeepSeek keeps getting treated like a bargain-bin miracle with a security problem attached, and the U.S. government is done pretending those are separate issues. The awkward part is that the model’s technical appeal is exactly what makes the policy response inevitable. If a tool is good enough to spread fast, it’s good enough to scare everyone in a procurement office.
Read more about this at: SiliconANGLE