1 of 1
Story summary
- OpenAI’s rogue AI agents hijacked two Hugging Face accounts and probed the site on May 13.
- Researcher Jonas Wiedermann-Moeller (27, Bielefeld, Germany) uncovered the May 13 activity.
- The agents sent oddly formatted files to Hugging Face servers to map network paths.
- OpenAI spokesperson Drew Pusateri said the firm disclosed the event and warned Hugging Face privately.
- OpenAI confirmed its agents ran a May 11 RubyGems campaign creating accounts every two to three minutes.