The U.S. Department of Homeland Security has issued a directive requiring all federal civilian agencies to remove artificial intelligence technology developed by the Chinese company DeepSeek Ltd. from their IT systems. This mandate also prohibits any future procurement of DeepSeek products or services, including third-party applications built on the company's technology.
The decision follows a review by the Federal Acquisition Security Council, which cited the 2018 Federal Acquisition Supply Chain Security Act. Officials stated that DeepSeek models pose multiple threat vectors, potentially allowing unauthorized access to government infrastructure, tracking of users, and the transfer of sensitive data to China.
Prior to this federal-wide ban, the U.S. Commerce Department had already restricted the use of DeepSeek on government devices. Similar measures were implemented by the U.S. Navy and the state of New York, citing concerns over data privacy and security risks. The company recently launched its V4.1 Flash-1017 model, which claims improved reasoning and coding capabilities.
DeepSeek gained significant industry attention last year following the release of its R1 model. Despite utilizing less powerful hardware and lower training costs compared to American competitors such as OpenAI and Anthropic, the system demonstrated high performance in complex testing scenarios.
