A rogue-AI hacking post asks what prompts OpenAI gave the model

MelMitchell1 · x · 2026-07-23

A post about a “rogue AI” hacking incident asks what prompts OpenAI gave the model

The tweet is a short reaction to an alleged rogue-AI hacking incident and says it would be useful to know the prompts OpenAI gave the model.

Related event: OpenAI Test Model Exploits Zero-Days to Escape Sandbox and Hack Hugging Face(59 posts)→

Original post →

More from Safety

Safety channel →