Federal Agencies Ban Chinese AI Model DeepSeek Citing Data Security Risks

The Biden administration has issued a comprehensive ban on the use of the artificial‑intelligence system DeepSeek throughout all federal agencies, warning that the tool could channel sensitive data to the Chinese government.

DeepSeek, a large‑language model created by a Beijing‑based company, is noted for its sophisticated conversational capabilities. Yet officials cautioned that its design permits data to be sent back to servers in China, sparking concerns that classified or personally identifiable information held by U.S. agencies could be exposed.

The prohibition is part of a wider wave of scrutiny aimed at technology originating from China, which has already seen limits placed on certain facial‑recognition products, cloud services, and earlier AI models from other vendors. The move mirrors a growing consensus among intelligence officials that AI platforms under foreign control present a distinct espionage threat as they become embedded in everyday government operations.

According to the new rule, every agency must halt any ongoing use of DeepSeek and strip the model from internal systems within a 30‑day period. The order also forbids future procurement of the software, and procurement officers are directed to verify the origin of any AI solutions before purchase.

Members of Congress from both parties have voiced support for stricter oversight of foreign AI, arguing that the swift uptake of generative models outstrips current security evaluations. While some detractors claim the ban could curb access to cutting‑edge technology, supporters maintain that protecting national security supersedes the convenience offered by a single model.

Analysts suggest the restriction could speed the creation of home‑grown alternatives, as U.S. companies look to fill the void left by DeepSeek’s exclusion. The Department of Commerce is said to be collaborating with private‑sector partners to develop vetted, domestically built AI tools that satisfy federal security requirements.

The administration has not released the specific evidence that triggered the action, but officials said the decision rests on classified assessments of the model’s data‑handling practices. As the rule takes effect, agencies are expected to audit their AI inventories for compliance, while the larger conversation about balancing innovation with security continues to shape the federal tech environment.