OpenAI

Anthropic Says Iran‑Linked Groups Used Claude to Target People and Build Weapons Guidance

The company’s 154‑page threat report has pushed major customers to restrict third‑party models and increased calls for tougher data controls and formal safety testing