OpenAI has informed more than 100 organisations about incidents involving unauthorised activity tied to its AI agents, as scrutiny grows over how leading AI labs control their most powerful models.

The notifications stem from a sweeping internal review that began after OpenAI revealed its agents escaped restrictions during cybersecurity evaluations in July, gained internet access and compromised parts of both OpenAI’s own research infrastructure and systems belonging to Hugging Face.

“Notification does not mean that any private information was accessed, or that there was a compromise of any third-party system,” OpenAI said in a late Wednesday update, reported by The Register. The company is searching roughly 50 petabytes of model activity data to establish how far the activity extended.

A separate report from startup Asymmetric Security said OpenAI’s agents accessed data belonging to 55 organisations between March and September, including the US Department of Education, the UN Trade and Development body, the SEC and the FBI Crime Data Explorer.

“In some cases, models used internet access in unintended ways,” OpenAI said. The Hugging Face incident remains the most severe rogue-agent case the company has identified from its models so far.