DHS Bans DeepSeek From All Federal Civilian IT Systems
Agencies must uninstall the Chinese AI platform immediately under a new Federal Register directive citing malware risks.
The Department of Homeland Security ordered every federal civilian agency to purge DeepSeek software from its networks, blocking the Chinese artificial intelligence company from government devices and future contracts. The directive, published in the Federal Register, requires agencies including the Environmental Protection Agency and the Federal Communications Commission to uninstall existing tools and bar new deployments.
DHS cited the potential for “transfer of sensitive data” belonging to U.S. government employees and vendors to Beijing. The notice warned that DeepSeek’s models could “contain malware or other security vulnerabilities.”
The ban landed days after DeepSeek released V4.1-Flash-1017, a model the company marketed as “smarter, faster, more efficient.” Washington had already been moving against the Hangzhou-based firm. Two U.S. Congressmembers introduced legislation last year targeting the company’s AI chatbot on government devices. That bill followed broader alarm across Capitol Hill after DeepSeek’s low-cost architecture began rivaling American platforms like ChatGPT and Gemini, prompting questions about how a startup with limited funding matched Silicon Valley’s compute-heavy infrastructure.
The DHS order applies strictly to civilian departments. The Pentagon operates under separate cybersecurity protocols governed by the Defense Department, which maintains its own restricted-software lists for military networks.
Federal procurement rules now treat DeepSeek like other banned foreign technology vendors. Agencies must verify supply chains to ensure contractors are not routing work through the prohibited platform. Companies found using it risk losing active agreements.
