Overview
- Caitlin Kalinowski said she quit because the company rushed into deploying its models on classified Defense Department networks without sufficient governance or clearly defined guardrails.
- OpenAI confirmed her departure and reiterated that its agreement includes explicit red lines prohibiting use for domestic surveillance or fully autonomous lethal systems, following recent contract revisions.
- Consumer and reputational fallout has grown, with app analytics showing a 295% spike in ChatGPT uninstalls and Anthropic’s Claude rising to No. 1 on the U.S. App Store as employees and researchers voice concerns.
- The Pentagon canceled Anthropic’s contract, labeled the company a supply‑chain risk after it refused terms allowing all lawful uses, and Anthropic says it will challenge the designation in court.
- A senior Pentagon official described heavy operational dependence on Anthropic’s Claude and said the department moved to add OpenAI and other vendors on similar terms to ensure redundancy in classified workflows.