US strips DeepSeek from government systems over data-security fears

The United States has taken its biggest step yet against Chinese AI company DeepSeek: ordering the entire federal government to remove it.

The Department of Homeland Security (DHS) has issued a directive requiring civilian federal agencies to strip DeepSeek's products and services from their computer systems and to stop buying them. The ban does not stop at DeepSeek's own software — it also covers third-party products and services built using its technology. Agencies can request special exemptions.

The reason is national security. The department says DeepSeek's models "present multiple threat vectors" and that their use "may result in the transfer of sensitive data" to China if used by government employees or vendors. It warns that sensitive data from US government users could potentially be accessible to Chinese authorities under Chinese law.

This is not the first move. In December 2025 the US passed a law targeting DeepSeek's use in intelligence and national-security systems; this new order extends the ban to all civilian agencies, covering both new purchases and existing systems. The Commerce Department blocked DeepSeek from government devices last year, the US Navy ordered all its members not to use the model, and the state of New York has banned it from government devices and networks over data-privacy and censorship concerns.

China has pushed back hard. Foreign Ministry spokesperson Mao Ning said Tuesday that Beijing opposes "overstretching the concept of national security and politicizing technology and trade issues," adding: "We will resolutely safeguard the legitimate rights and interests of Chinese enterprises."

The dispute is part of the wider US–China technology rivalry. DeepSeek, founded by Liang Wenfeng, shook the AI industry in January 2025 when it released its R1 reasoning model — a low-cost model that sent US tech stocks into a tailspin. The US has also accused DeepSeek and five other Chinese AI companies — including Alibaba and Moonshot AI — of using "distillation" to siphon knowledge at industrial scale from American AI systems, according to a September report by the cybersecurity agency CISA. OpenAI told US lawmakers earlier this year that it had observed suspected attempts by DeepSeek to copy its models, and Anthropic said it had identified unauthorized attempts by DeepSeek and other Chinese labs to collect outputs from Claude.

Meanwhile, reports say China is considering export restrictions on 23 AI-related technologies, including chip designs and certain advanced AI models.

What does this mean for you?

Small businesses: Many cheap business tools are built on models like DeepSeek. This story is a reminder to ask the key question: when you pour customer data into a free or cheap tool, where does that data go — and whose laws decide who can reach it? If you handle customer, business, or payment information, know the privacy policy of the tool you use.

Developers: The ban covers services built on top of DeepSeek — but its models are open-weight. Developers can download a model and run it on their own servers, where the data never leaves their systems. That is the crucial difference between using the API and running it yourself.

Students and everyday users: DeepSeek is hugely popular across East Africa because it is free and fast. This ban does not affect you — it covers US federal agencies. But it is a general lesson: never paste your ID number, bank account details, or business secrets into any chatbot, free or paid.