The US Department of Homeland Security (DHS) has issued a directive ordering all federal civilian agencies to stop using artificial intelligence tools developed by Chinese startup DeepSeek. The notice requires agencies to immediately remove DeepSeek products, as well as third-party software built using DeepSeek technologies, from all government IT devices.
According to an accompanying DHS document, DeepSeek presents multiple threat vectors due to its servers being located in China. US officials cited Chinese legal frameworks that give the state broad authority to access private corporate infrastructure, warning that utilizing the AI models could expose sensitive US government data to foreign access. The directive applies to civilian agencies, while military and intelligence services operate under separate jurisdictions.
The ban targets one of China’s prominent AI developers, which gained widespread industry attention in early 2025 by offering high-performance, low-cost models. The policy escalation comes amid accusations from US officials that Chinese firms steal model capabilities, despite planned bilateral talks on AI incident communication channels between US and Chinese leadership.
Why it matters
Excludes DeepSeek and derivative models from US federal government procurement and enterprise supply chains.
Forces enterprise developers servicing government contracts to audit software dependencies for Chinese AI models.
Heightens geopolitical compliance risks for international cross-border AI deployments.
Source: theguardian.com



