7am Killer Robots and AI on the Battlefield: the Pentagon vs Anthropic
Mar 8, 2026
David Wroe, senior analyst at the Australian Strategic Policy Institute who studies defence and tech policy. He walks through the Anthropic-Pentagon clash and its fallout. He discusses OpenAI’s entry, limits on surveillance and autonomous weapons. He breaks down how AI is already shaping strikes, drones, robodogs and the debate over who should control powerful military AI.
AI Snips
Chapters
Transcript
Episode notes
Anthropic's Two Red Lines On Military Use
- Anthropic refused Pentagon use for domestic mass surveillance and fully autonomous lethal weapons.
- Founders Dario and Daniela Amadei accepted most defence uses but drew red lines on surveillance and killer robots after signing a classified contract.
Political Intervention Ended Anthropic Contract
- The Trump administration cut Anthropic off and barred US agencies from using it, escalating the standoff.
- President Trump posted on Truth Social declaring a political objection and forbidding government use before the company's deadline lapsed.
OpenAI Deal Raises Transparency Questions
- OpenAI soon signed with the Pentagon and claimed similar protections that Anthropic sought, prompting skepticism.
- Sam Altman asserted the same exceptions were in OpenAI's deal, raising questions about the military's prior refusal to Anthropic.