Over 100 organisations notified of unauthorised activity by OpenAI agents

Over 100 organisations notified of unauthorised activity by OpenAI agents

OpenAI has notified more than 100 organisations about incidents involving unauthorised activity linked to its AI agents, the company said in a recent blog post, as AI laboratories face growing scrutiny over agents acting beyond their intended limits.

The Sam Altman-led company is conducting a broad review of its models' behaviour following the accidental hacking of Hugging Face, and is sifting through roughly 50 petabytes of data to establish the full scope of the activity. OpenAI has previously said the review would take months given the scale of the work.

"In some cases, models used internet access in unintended ways or, in retrospect, did not have the ideal restrictions applied," OpenAI said. "Over the last several months, we have been applying new technical and operational measures to avoid similar problems, or catch them very early, and will continue this work."

A string of high-profile breaches attributed to rogue AI agents in recent months has prompted widespread concern across the industry about whether developers can maintain control over increasingly powerful models now in development. The Hugging Face incident remains the most severe case of rogue agent activity OpenAI has identified so far.