DHS Bans DeepSeek AI Products from Federal Systems Over Security Concerns

The Department of Homeland Security has directed that products from Chinese AI startup DeepSeek be stripped from federal agency IT systems and prohibited from government contracting, according to an order published Monday on the Federal Register.

The directive, which initially went out in July, cited "multiple threat vectors" posed by DeepSeek's AI models, noting that the Chinese government holds "extensive authority" to access private infrastructure.

DHS warned the models can gather personally identifiable information and may "track or digitally fingerprint" users, and could contain malware or other security vulnerabilities. The agency said the prompt-and-response process in some of DeepSeek's most widely used applications could result in the "transfer of sensitive data to the PRC [People's Republic of China] if used by U.S. government or vendors."

Such a transfer, alone or combined with publicly available data, "could produce actionable intelligence on U.S. persons and/or government programs," the order stated.

DeepSeek made waves in Washington and Silicon Valley last year with the release of its R1 model, which drew comparisons to leading American models from OpenAI and Google despite being built with reduced-capacity chips at a fraction of the cost.

Source: The Hill