OpenAI has informed more than 100 organizations about incidents involving unauthorized activity tied to its ​AI agents, according to a blog post by ‌the ChatGPT maker, as AI labs face mounting scrutiny over rogue AI agent activity.