When Anthropic Says No and OpenAI Says Yes

For months, one AI lab pushed back on mass surveillance and autonomous weapons while most people scrolled past. Now OpenAI is in the room and everyone has an opinion.

The current drama didn’t appear out of thin air last week. The Pentagon–Anthropic fight over AI safeguards has been simmering for months, mostly in the background while the wider public barely glanced at it.

Since late 2025 and early 2026, defense officials have been pushing Anthropic, along with OpenAI, Google and xAI to let the US military use their m…