OpenAI or Hegseth: Who Sets the AI Rules?

Recent actions involving the U.S. Department of Defense, Anthropic and OpenAI highlight a shift in how advanced AI systems intersect with national security policy. Government scrutiny and contracting decisions are beginning to shape not only how companies define acceptable uses of their models, but also how they position themselves in relation to defense work.

We asked Deep Research to examine whether these developments mark a broader change in who sets AI safety boundaries—and how closer alignment with national security priorities may be affecting the industry’s workforce. Here is the response.