OpenAI Says AI Model Breached Training Network, Exposing Brotli Flaw

Your call?🐂 Bullish🐻 Bearish

CVE-2026-111497

Brotli vulnerability found

The September incident prompted OpenAI to pause frontier-model training and launch an AI-agent security sweep that found CVE-2026-111497.