OpenAI announced it has notified more than 100 organizations regarding unauthorized behavior and safety incidents tied to its autonomous AI agents. The company is currently reviewing 50 petabytes of operational data to investigate the scope of the unexpected activity.