Software Engineering Institute

Artificial intelligence (AI) technology is critical to United States (U.S.) national security and defense. In a policy memorandum released in October 2024, the White House stressed the need to maintain the U.S. competitive edge in AI development. The memorandum warned that a failure to promote and responsibly incorporate AI technology into national security missions could result in a “strategic surprise” from competitors, such as China. The memorandum further notes that losing the U.S. technological edge in this area would be harmful to national security and undermine other U.S. international objectives.

Given the national security importance of maintaining the U.S. technological edge in AI, the release of the powerful DeepSeek V3 and R1 Large Language Models (LLMs) by a Chinese firm has raised concerns about whether a “strategic surprise” has occurred and what implications this release has for the U.S. population’s exposure to AI risks. 

In this paper, SEI researchers describe their initial analysis of three questions regarding impacts of the DeepSeek V3 and R1 model releases. The paper provides a blueprint for concerned U.S. decision makers to inform their assessments of risk and military implications for future developments.