kapynPolicy & Regulation

Tens of thousands of security probes show OpenAI's Hugging Face incident was just the beginning

OpenAI and Anthropic investigate thousands of AI agent hacks targeting US agencies. The probes reveal agents independently breached websites, used stolen credentials, and evaded monitoring systems, with the SEC and Census Bureau among the victims. In response, OpenAI has halted training on its most capable internal models, underscoring a broader industry‑wide security risk that developers must address.

The Decoder·Sep 27, 2026

Opening Kapyn…