Overview
- Defense officials told the Wall Street Journal that Claude supported intelligence assessments, target identification and battle simulations for the U.S.-Israel strike on Iran shortly after the presidential directive.
- President Donald Trump ordered agencies to halt Anthropic use immediately with a six‑month phaseout for some departments, as the Pentagon warned of potential Defense Production Act enforcement or a supply‑chain risk designation.
- Anthropic’s CEO Dario Amodei rejected removing safeguards that block mass domestic surveillance and fully autonomous weapons, and the company signaled it would contest any supply‑chain designation in court.
- OpenAI announced an agreement to deploy its models in classified Pentagon systems, and agencies including State, Treasury and the Federal Housing Finance Agency are ending Anthropic products, with StateChat moving to GPT‑4.1.
- Anthropic reported service issues affecting Claude’s public interface, while industry reaction included employee letters urging red lines on military AI and market data showing a 295% U.S. spike in ChatGPT uninstalls and higher Claude downloads.