US bans agencies from using Chinese AI DeepSeek, warns sensitive data may go to China

The AI race just saw another development. The US now wants federal agencies to stop using AI models made by Chinese AI lab DeepSeek. In a notice sent on Monday, the US Department of Homeland Security (DHS) has ordered all federal civilian agencies to remove DeepSeek’s products and services from their IT systems. The notice warns that using DeepSeek’s AI models may lead to a transfer of data to China.The DHS sent the directive to civilian agencies. Apart from using DeepSeek’s products, the notice bars the agencies from acquiring the products or services from the Chinese AI lab. It also covers products and services developed by third parties using DeepSeek models.Though, the order does not apply to national security systems or secure compartmented information systems, including those overseen by the Department of Defense and the intelligence community.Why is the US banning DeepSeek?As per the DHS, DeepSeek models presented "multiple threat vectors.” It warned that their use "may result in the transfer of sensitive data” to China if used by US government employees or vendors. The department said DeepSeek's servers were located in the People's Republic of China, whose legal system allowed government cyber and intelligence authorities extensive access to private company infrastructure. It also said DeepSeek's account creation process collected personally identifiable information and could track and digitally fingerprint users.According to the notice, transferring such information to China, either independently or alongside publicly available data, "could produce actionable intelligence on US individual and/or government programmes." It added, "The models themselves may contain malware or other security vulnerabilities." Agencies must remove DeepSeek products, products from its affiliates and third-party services developed using its technology. They must also exclude them from approved product lists, service-provider equipment lists, e-commerce markets and other procurement channels.Though an agency may request an exception in writing within 60 days, citing compelling reasons such as effects on mission-critical functions or national security considerations. Until the products are fully removed, agencies have been advised to monitor incoming and outgoing traffic, strengthen logging and data-exfiltration controls, and flag unusual activity.This is not the first time a US body has cracked down on DeepSeek. The US Department of Commerce and the US Navy previously banned DeepSeek on government devices. The states of New York, Tennessee and Virginia have also taken similar steps.DeepSeek has grown in popularity since its public release in 2025. Its open-source models are less expensive to use than closed models offered by US developers, while providing similar capabilities. US companies including Snowflake, Vercel and OpenRouter have incorporated DeepSeek into services offered to customers.The directive comes amid wider Trump administration action against Chinese tech companies. When talks about a slow down in frontier AI development gained traction, US President Donald Trump claimed that China would be the biggest benefactor of such a move in the US. “There is a sick conspiracy going on against AI and Data Centers, and the only one that is happy about it is China,” he wrote on Truth Social last month.US AI labs such as Anthropic and OpenAI have also accused Chinese players, including DeepSeek, of distilling their AI models. Distillation refers to the process of using a larger model’s reasoning and output to train a smaller model.- Ends