DeepSeek banned from U.S. federal civilian agencies under new DHS order

Civilian federal agencies must stop using DeepSeek AI products and remove them from their systems.

The Department of Homeland Security (DHS) has banned the use of Chinese artificial intelligence model DeepSeek and its affiliates by all civilian federal agencies. This directive, published in the Federal Register and effective immediately, cites national security concerns under the Federal Acquisition Supply Chain Act. Secretary of Homeland Security Markwayne Mullin stated that DeepSeek models pose multiple threat vectors to federal IT systems, including data transmission to servers in China, potential malware, and risks to privacy and mass data collection. Mullin highlighted that Chinese national security laws could compel domestic companies to share data with the government, potentially leading to actionable intelligence on U.S. persons and programs. The ban requires agencies to cease procurement of DeepSeek products and remove any existing installations, with exceptions possible only under strict, pre-approved justifications. Agencies are also advised to enhance security controls to mitigate risks from prior use of the banned AI tool.

The DHS ban on DeepSeek signifies a more aggressive stance by the U.S. government in scrutinizing and restricting foreign AI technologies, particularly those originating from China. This move suggests that national security concerns, including data sovereignty and potential state-sponsored espionage, are now overriding the technical capabilities or cost-effectiveness of AI models. It sets a precedent for how other federal agencies and potentially even private sector entities might approach the procurement and integration of AI tools from geopolitical rivals.

This action directly impacts the accessibility of advanced AI models for U.S. government operations, potentially creating a compliance burden and requiring agencies to seek alternative, domestically sourced or vetted AI solutions. The requirement for agencies to submit exception requests with compelling justifications indicates a high threshold for any continued use, signaling a clear preference for minimizing exposure to perceived risks associated with foreign AI. The DHS is effectively drawing a line, prioritizing security over the broader adoption of cutting-edge AI from specific foreign entities.