US orders federal agencies to ban DeepSeek AI tools

The US Department of Homeland Security has issued a directive ordering all federal civilian agencies to purge AI technology developed by Chinese laboratory DeepSeek from their systems. The mandate applies to direct deployments of DeepSeek models as well as third-party services built on top of the company's architecture. Officials cited national security risks associated with the use of Chinese-origin software across civilian federal infrastructure.

The ban comes amid tightening scrutiny over the deployment of foreign AI models within public administration and government systems. DeepSeek, which produces open-weight and proprietary models, had gained traction across developers and organizations due to competitive benchmark results and lower API pricing. Federal administrators must identify and remove any operational instances of the technology across internal networks and partner workflows.