Executive Summary
The recent emergence of DeepSeek AI sparked global concerns over its security implications, vulnerabilities, and potential for influence operations due to its Chinese affiliation. Multiple nations, including Italy and Taiwan, have limited or banned its use, citing issues of data and intelligence security. This happened following a large data leak.
Experts and users point out that DeepSeek responses align with Chinese government interests and narratives, and comparative tests of the accuracy of the responses had a high fail rate. Officials believe that the AI model could enable disinformation and propaganda on a large scale, much like other AI platforms.
Despite these concerns and challenges, we assess that DeepSeek’s novel minimalist-resource model will upset global competition in the AI arena. DeepSeek’s rise also signals an intensification in China-US technological competition, especially considering the launch of the US “Stargate” project. Continual monitoring of DeepSeek as an innovation challenge and potential strategic threat will remain an urgent and relevant matter in the coming months, if not years.
