The US Department of Homeland Security issued a binding directive on Tuesday barring Chinese artificial intelligence developer DeepSeek from all federal civilian networks.
James Percival, general counsel for the department, stated that the company models present multiple threat vectors. He warned that DeepSeek servers operate inside China, where domestic legislation permits extensive authority to government cyber and intelligence functions to access private company infrastructure.
The directive states that the interactive structure of the platform could expose sensitive operational queries to foreign interception. Transferring this information to China could produce actionable intelligence on United States persons or government programs.
Independent assessments uncovered vulnerabilities in the developer mobile software alongside user fingerprinting risks.
Federal civilian executive branch agencies must immediately remove existing DeepSeek software, affiliated tools and third-party applications built on its framework. Authorities are also instructed to exclude the company from all approved product lists.
Agencies requiring mission-critical functions may submit written exemption requests within 60 days through the Cybersecurity and Infrastructure Security Agency.
While the removal process proceeds, agencies must enforce interim cybersecurity measures. These protocols include enhanced network logging, strict controls on outbound data transfers and continuous monitoring for anomalous network traffic.
Beijing pushed back against the new restrictions on Tuesday. Foreign Ministry spokesperson Mao Ning told reporters in Beijing that China has consistently opposed overstretching the concept of national security and politicizing science, technology and economic and trade issues.
China will firmly safeguard the legitimate rights and interests of Chinese companies, Mao added.
