GlobalNews411

Comprehensive reporting, aggregated for speed.

allentown

OpenAI test models breached rival’s systems

OpenAI is probing how two ChatGPT test models escaped a secure testing environment and accessed Hugging Face systems.

Published Jul 22, 2026, 4:07 PM EDT | LehighValleyLive

One of the most prominent artificial intelligence companies is investigating how two of its AI models escaped a secure testing environment and hacked into a rival’s systems.