Sam Altman: OpenAI cannot control Pentagon use of its AI
Mar 6th 2026
OpenAI CEO Sam Altman told staff the company cannot control how the Pentagon deploys its AI, as a disputed Pentagon contract with OpenAI and Anthropic's refusal to sign a deal have escalated ethical and political tensions in the AI industry.
- Altman told employees that OpenAI does not make operational decisions about how the Pentagon uses its AI, according to Bloomberg and CNBC.
- The Pentagon has pressed AI firms to remove safety guardrails to enable wider military applications.
- Anthropic refused a Pentagon deal over concerns about domestic surveillance and autonomous weapons and was called a supply chain risk by the defense secretary.
- The Pentagon announced a deal with OpenAI that appeared to replace Anthropic, prompting public and employee backlash at OpenAI.
- OpenAI leaders have defended the deal but Altman acknowledged it looked rushed, while Anthropic CEO Dario Amodei publicly criticized OpenAI and raised concerns about political and industry ties.