US Orders Removal of China's DeepSeek AI From Federal Government Systems

The US government has moved from warning about DeepSeek to ordering it out. According to Reuters, the Department of Homeland Security is ordering the removal of DeepSeek products, and third-party services that use the Chinese AI model, from civilian agency IT systems, according to a notice published Monday. It also bans civilian agencies from acquiring them.

DHS said DeepSeek models pose "multiple threat vectors" and warned that use could "result in the transfer of sensitive data to the PRC" if used by government employees or vendors. The notice relies on the Federal Acquisition Supply Chain Security Act of 2018, and OfficeChai reports it is the first time that law has been used against an AI company. Several details define how far the order goes:

• It reaches third-party products built using DeepSeek, not only DeepSeek's own tools, so vendors serving agencies may need to audit what is inside their software
• It applies to civilian agencies except national security systems and sensitive compartmented information systems, which fall under the Pentagon and the Director of National Intelligence
• DHS said the transfer of data, alone or combined with public data, could produce actionable intelligence on US persons or government programs

What the order does not do is just as important. It does not ban DeepSeek for American companies or individuals, and reporting notes it does not cover state governments or foreign governments.

The Chinese embassy did not immediately comment to Reuters. The practical impact falls on agencies that experimented with DeepSeek and on contractors who embedded its models in products sold to the government, who now face an audit and possible contract questions.

The order also fits a wider pattern: Washington has increasingly treated Chinese AI models as a supply-chain security issue, and a procurement rule is the easiest lever to pull because it can take effect without new legislation.