The Pentagon is facing fresh scrutiny after reports emerged that its personnel have been using an early version of the Chinese generative artificial intelligence model DeepSeek for months. The revelation, first reported by Bloomberg, has raised serious national security concerns, with US defense officials scrambling to assess the potential risks associated with the technology.

According to sources cited by Bloomberg, staff within the Department of Defense (DoD) have been downloading and utilizing an earlier version of DeepSeek on their workstations since the fall of 2024. It was only after a recent surge in the AI model’s popularity that the Defense Information Systems Agency (DISA), which manages the Pentagon’s IT networks, moved to partially block its usage.

The timing of this revelation is particularly alarming for US defense and intelligence agencies. DeepSeek, developed by a Chinese startup, has positioned itself as a competitive alternative to Western AI models, including OpenAI’s flagship products. Earlier this month, the company released DeepSeek R1, an open-source AI model that reportedly outperforms some of its US counterparts. The accessibility of this model-allowing users to download and run it on their own servers for free-has sparked concerns about national security and corporate competition in the AI sector.

Following the discovery that multiple defense employees had used DeepSeek’s latest chatbot version for at least two days, DISA blocked access to the company’s website on January 28. However, efforts to contain the AI’s spread have been complicated by the fact that thousands of Pentagon employees continue to access DeepSeek via the web platform Ask Sage. This platform, which offers access to multiple AI models, hosts DeepSeek on US-based servers, effectively bypassing direct restrictions.

Pentagon officials are now conducting an internal audit to determine the extent........

© Blitz