DeepSeek AI ban: US orders agencies to drop Chinese startup's tools
The US government on Monday announced a DeepSeek AI ban for federal civilian agencies, in a directive from the Department of Homeland Security that also covers third-party tools built with the Chinese startup's products.
DeepSeek AI ban: US orders agencies to drop Chinese startup's tools
What the ban covers
A notice issued by the Department of Homeland Security on Monday announced the DeepSeek AI ban on products developed by the company, along with tools created using its technology by third parties. Agencies are required to remove them from their IT devices. The directive applies to civilian government agencies.
Why the department is concerned
The department laid out several security risks in a document attached to the notice. According to DHS, the company's models carry dangers it described bluntly.
DeepSeek's models present multiple threat vectors.
The department's stated concerns include:
- Use of DeepSeek AI tools may result in the transfer of sensitive data to China if used by US government employees or vendors.
- The company's servers are located in China, where the legal system permits the government extensive authority to access private company infrastructure.
Limits of the order
The Department of Homeland Security can issue direction to civilian agencies but not to the military or intelligence services, leaving their use of the technology outside the ban's scope. The US defense department did not respond to an AFP request for comment, and DeepSeek has been contacted for comment.
DeepSeek's rise and the US-China backdrop
DeepSeek, one of China's top AI labs, shook up the industry in January 2025 with a low-cost, high-performance chatbot that challenged assumptions of US dominance. Its V4 model arrived in April, following an October update that produced a faster, more lightweight version of V4.
Washington's accusations against the firm centre on claims that DeepSeek and other leading Chinese AI companies have systematically stolen the capabilities of cutting-edge models built by American rivals. When Xi Jinping and Donald Trump met at the White House in September, China and the US agreed to set up a communication channel for AI incidents. The subject is expected to arise again at a summit between the two leaders scheduled for November in the Chinese city of Shenzhen.
