# US DHS Bans DeepSeek AI Across Federal Civilian Agencies New directive mandates removal of all DeepSeek-related technology from federal civilian infrastructure. - **Source:** [www.theregister.com - Articles (theregister.co.uk)](https://brifing.si/source/e6d4ec13-4d0e-407a-b96d-39609d76c3b1) - **Published:** 2026-10-20 - **Topic:** [Business](https://brifing.si/c/business) - **Original article:** - **On Brifing:** https://brifing.si/s/a91b5291-9bb7-4fd4-9590-1e7978d9ad40 ## DHS Issues Sweeping Ban The US Department of Homeland Security issued a directive requiring all federal civilian agencies to remove DeepSeek products. This order explicitly targets DeepSeek-affiliate technologies and any third-party services built upon their implementations. The move follows months of escalating scrutiny regarding the Chinese AI lab's potential risks to national security and data integrity. ## Ambiguity in Implementation The directive remains vague regarding the definition of products developed using DeepSeek. It is unclear if the ban covers software coded with AI assistance or services accessing models via APIs. This creates significant compliance challenges for agencies, as they must now identify hidden dependencies within their existing software supply chain architectures. ## Gateway Service Risks Federal agencies face inadvertent exposure through AI gateways like OpenRouter or Vercel. These platforms automatically route requests to the best available providers, potentially utilizing DeepSeek models without explicit user intent. Agencies must now implement strict controls to ensure their AI traffic does not bypass these new security mandates through automated model selection. ## Legislative and Defense Context The ban follows earlier actions by the US Defense Information Systems Agency, which blocked DeepSeek on Pentagon networks in January 2025. Representatives Darin LaHood and Josh Gottheimer previously introduced the No DeepSeek on Government Devices Act. While that bill stalled in committee, various government branches independently moved to exile the technology. ## Market Impact and Competition This policy shift may benefit American AI companies by eliminating low-cost, high-capability competition from Chinese models. Agencies restricted from using open-weight models like those from DeepSeek will likely shift toward domestic providers. This transition could lead to higher operational costs for taxpayers as agencies move away from cheaper, international AI alternatives. ## Mobile App Security Concerns Reports from December 2025 highlighted severe vulnerabilities in DeepSeek’s mobile applications, including plaintext data leaks and insecure server connections. While these issues mirror common problems found in many third-party apps, the specific risks involve the ingestion of sensitive government prompts. These technical shortcomings have fueled the broader narrative regarding potential data capture. ## Expert Critique of Scope Katie Moussouris of Luta Security argues the order is overly broad and practically unenforceable. She notes that while banning the app is logical, prohibiting any software developed using DeepSeek is technically impossible to verify. Current software bills of materials do not track the specific AI tools used by developers during the coding process. ## Vulnerability and Future Outlook NIST findings suggest DeepSeek models are more susceptible to prompt injection and jailbreaking than US counterparts. However, critics argue that applying these restrictions to US-hosted versions of the models ignores the mitigation of data flow risks. The order sets a precedent for how federal agencies will manage third-party AI dependencies moving forward. ## Sources covering this story - https://www.theregister.com/security/2026/10/20/deepseek-deep-sixed-at-all-us-federal-civilian-agencies/5304010 — --- _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._