OpenAI has agreed to deploy advanced AI systems in classified U.S. military environments, sharply expanding its Pentagon collaboration. This deal came immediately after the Trump administration blacklisted competitor Anthropic, labeling it a national security risk and banning federal use of its technology. OpenAI’s agreement is under scrutiny due to contract language that grants the Pentagon use of its AI for “all lawful purposes,” mirroring language Anthropic refused. While OpenAI claims to maintain non-negotiable red lines—such as prohibitions on mass domestic surveillance, independent autonomous weapons, and high-stakes automated decisions—these restrictions are tied to existing government policies and legal frameworks, not independent standards. Critics argue this approach lets government-defined legality, not company-set ethics, determine AI use, raising concerns about enforcement and interpretation in sensitive national security contexts. The announcement triggered backlash, with the “QuitGPT” movement organizing subscription cancellations and driving Anthropic’s app ahead of ChatGPT in downloads. The episode spotlights philosophical differences: OpenAI relies on legal and technical safeguards, while Anthropic insisted on explicit, binding contract limits. Whether these differences result in meaningful real-world protections remains debated.