Trump Administration Bans DeepSeek AI Across Federal Agencies
Homeland Security mandates removal of Chinese AI models citing data privacy and security risks.
1Federal Ban Implementation
The Department of Homeland Security issued a directive requiring federal civilian agencies to remove all DeepSeek technology. This mandate extends to third-party services relying on DeepSeek models. Secretary Markwayne Mullin emphasized that the software poses significant risks, including potential malware integration and the unauthorized collection of sensitive personally identifiable information from government users.
2Data Privacy Concerns
Officials are concerned that DeepSeek’s account generation process facilitates digital fingerprinting and data harvesting. The administration fears this information could be transferred to the Chinese government, creating a direct national security vulnerability. Consequently, agencies like the Department of State must purge these tools to prevent potential foreign intelligence exploitation of federal infrastructure.
3Market Disruption History
DeepSeek caused significant industry volatility last year following the release of its R1 model. The model demonstrated high capabilities at a fraction of the cost compared to American counterparts, triggering a trillion-dollar stock market wipeout. Previously, the U.S. Navy and the Department of Commerce had already restricted personnel from utilizing these specific AI tools.
4The Distillation Controversy
American AI giants, including Anthropic, have accused DeepSeek of illicitly using distillation to train models. This process involves using outputs from advanced teacher models to train student models, which critics label as intellectual property theft. The U.S. government now views these adversarial training campaigns as a form of industrial-scale cheating against domestic firms.
5Escalating Security Warnings
A joint letter from the NSA, CISA, and the FBI recently identified DeepSeek as a primary actor in industrial-scale distillation campaigns. Treasury Secretary Scott Bessent has threatened potential sanctions against the company, characterizing these actions as clear IP theft. This aggressive stance reflects a broader government effort to protect American technological dominance from foreign interference.
6Geopolitical AI Competition
While American models currently lead global development, the gap is narrowing rapidly. The emergence of R1 and Moonshot’s Kimi K3 has intensified fears regarding the AI race. Despite promises of cooperation through the US-China AI Dialogue, Washington hawks continue to warn that Chinese advancements could provide Beijing with a decisive military advantage.
7Beijing’s Regulatory Stance
Interestingly, Beijing is also expressing concern regarding the security implications of its own AI sector. Authorities have proposed measures to implement export controls on the overseas deployment of certain Chinese models. This suggests that even the Chinese government is attempting to keep firms like DeepSeek on a tighter leash to manage domestic risks.
8Strategic Policy Outlook
The timing of this ban coincides with upcoming bilateral talks on AI security between the U.S. and China. Treasury Secretary Scott Bessent indicated that further actions are expected in the coming weeks. Ultimately, the administration is prioritizing defensive posture, arguing that offensive AI investment is ineffective if federal systems remain vulnerable to foreign-developed software.
Original article: Gizmodo — Trump Administration Bans DeepSeek AI Across Federal Agencies