China defends DeepSeek after US Homeland Security orders its AI stripped from federal systems over data fears

The Department of Homeland Security warned that DeepSeek's models could send sensitive data to Beijing and may contain malware. It said government use 'could produce actionable intelligence' on Americans. The ban quietly took effect at the end of July but was only made public on Monday. It covers the Treasury, State and Justice departments, while the Pentagon's agencies are excluded.

The Commerce Department, the Navy, New York, Tennessee and Virginia had already banned DeepSeek.

Anthropic and OpenAI claim Chinese firms including DeepSeek essentially copied their AI models.

The FCC votes on October 29 on barring Chinese labs, which tested 82% of US electronics last year.