CNN 5 Things One Thing: The Pentagon vs. AI Power Players: Did Anyone Win?
Mar 8, 2026
Dean Ball, a White House AI planner turned policy fellow, and Hadas Gold, CNN AI reporter, unpack the Pentagon’s pressure on AI firms and OpenAI’s controversial deal. They discuss Anthropic’s refusal over weapons and surveillance. The conversation covers supply‑chain ultimatums, political influence on procurement, and the shifting guardrails shaping AI’s future.
AI Snips
Chapters
Transcript
Episode notes
Pentagon Pressured Anthropic Over Red Lines
- The Pentagon demanded Anthropic allow use of its AI for "all lawful uses" and threatened a supply chain risk designation if it refused.
- Anthropic held two red lines: no autonomous lethal weapons and no mass domestic surveillance, citing safety and legal gaps.
OpenAI Agreed After Saying It Shared Anthropic's Stance
- OpenAI publicly said it shared Anthropic's red lines but then quietly signed a Pentagon contract late Friday, claiming those red lines were preserved.
- OpenAI employees felt the deal was rushed and opportunistic, prompting Sam Altman to apologize internally.
Politics Shaped Which AI Firms The Pentagon Embraced
- OpenAI's contract allows use for "all lawful purposes" but explicitly prohibits domestic surveillance and directing autonomous weapons, per the company.
- The Pentagon's willingness to accept OpenAI but punish Anthropic suggested political and personality factors.