Pentagon Pushes AI Firms To Enable Classified Use

WASHINGTON, Feb. 12, 2026 — The Pentagon is pressuring leading AI firms, including OpenAI and Anthropic, to make their generative models usable on classified networks without many of the companies' standard restrictions, officials told Reuters. The move, discussed at a White House event, follows an unclassified OpenAI deal and raises operational, safety, and policy concerns about deploying LLMs for mission planning and weapons-related tasks.