OpenAI CEO Sam Altman publicly defended the company’s newly inked Pentagon agreement, saying the deal aligns with OpenAI’s safety principles while enabling its AI to run on classified networks. The comments followed President Donald Trump’s directive for federal agencies to phase out rival Anthropic, which refused to permit “all lawful purposes,” including mass domestic surveillance and fully autonomous weapons. Pentagon officials labeled Anthropic a supply‑chain risk, escalating a broader fight over the rules for military AI. Altman said OpenAI moved quickly to “de‑escalate” tensions and negotiated terms he says could be extended to other labs, while acknowledging potential legal and political risks ahead. The episode underscores intensifying competition among AI firms for defense work as Washington seeks greater access to advanced models under clearer guardrails.
Related articles:
— Project Maven
— Joint Enterprise Defense Infrastructure (JEDI)
— NIST AI Risk Management Framework
— Google AI Principles
— Lethal autonomous weapon
