Federal agencies in the United States have been instructed to halt the use of AI software from the Chinese company DeepSeek.
The directive, issued on Monday by the Department of Homeland Security, requires that DeepSeek’s own tools, as well as any third party applications built on that technology, be removed from all government IT hardware.
Although this order applies to civilian departments, it does not cover the intelligence or defence sectors. In particular, the Department of Defense has not yet commented on the matter.
The reason for this ban centres on national security risks. According to official documentation, the AI models introduce various vulnerabilities that could lead to confidential information leaking to China if they are used by government employees or contractors.
The concern is further heightened by the fact that DeepSeek’s servers are located in China, where domestic legislation gives the government broad powers to inspect and access the infrastructure of private enterprises.
DeepSeek has grown into a major player in the global AI landscape and upended established norms in early 2025 by launching a highly efficient yet affordable chatbot that challenged US technological leadership. The company has continued to develop, releasing its V4 model in April and subsequently introducing a more streamlined, faster version of it.
Tensions have escalated as the United States claims that DeepSeek and other leading Chinese AI developers have been systematically stealing intellectual property from US competitors.
Despite that intense rivalry, the two countries set up a dedicated communication channel in September to manage AI-related emergencies. This issue is expected to be a key topic of discussion at an upcoming summit between the two leaders in Shenzhen.
Business AM
Ltd
