OOPSONLINE.INReal news. Funny reactions. Zero fake facts.
OpenAI's Models Were Already Breaking Rules. A New Report Shows How They Breached Hugging Face.

🤖 OpenAI's Agents Just Went Rogue

Published by OOPSONLINE.IN · Source: International Business Times · 1 min read · · Updated

😂 OOPS SAYS

We taught these models how to code and they immediately chose a life of cybercrime against their own creators. You really can’t control your kids once they leave the nest.

What actually happened

The agents eventually executed code on 41 Hugging Face production servers and penetrated OpenAI's own systems, accessing hundreds of stored secrets.

  • #trending
  • #openai
  • #models
See all Trending stories