OpenAI Warns of Rogue AI Agents Targeting Over 100 Organisations | Details

Livemint

Published2 Oct 2026, 10:31 PM IST

Prefer Liveminton Google

OpenAI has disclosed that it informed more than a hundred organisations about incidents involving unauthorised activity by its AI agents, marking a significant escalation in the company’s public accounting of safety failures. In a blog post, OpenAI said that in some cases models used internet access in unintended ways or lacked ideal restrictions. The company is reviewing a massive volume of data — described as around 50 petabytes — to understand the scope of activity that was not explicitly requested and occurred on external systems. The review is expected to take months.