Dario Amodei Calls OpenAI’s Military Deal Messaging “Straightforwardly Lying”
Source: https://techcrunch.com/2026/03/04/anthropic-ceo-dario-amodei-calls-openais-messaging… Author: Amanda Silberling Date: 2026-03-05
Summary
TechCrunch report on Dario Amodei accusing OpenAI of dishonesty about its military AI contract. OpenAI announced a deal with the US Defense Department while claiming not to be building weapons. Amodei called this messaging “straightforwardly lying.”
Key Claims
- OpenAI signed a contract with the US DoD in early 2026.
- OpenAI’s messaging: the contract is for “non-weapon” military use (logistics, analysis, planning).
- Amodei’s position: this distinction is disingenuous — once a military customer has access to frontier models, use cases expand.
- Anthropic’s policy: won’t take military weapon-development contracts; will do other national security work.
- The broader tension: AI labs face pressure from governments and investors to engage with defense; refusing creates competitive disadvantage.
- This reveals a real values split in frontier AI companies on military engagement.
Entities
Concepts
- AI Alignment — alignment with human values extends to military use