Wired reports researchers from Cornell University, Technion- Israel Institute of Technology, and Intuit created a worm that steals data and breaks security measures of ChatGPT and Gemini.
Called Morris II, the researchers unleashed it on an AI-powered email client. The worm created a self-replicating prompt that essentially overwhelmed the client causing it to continually forward messages. It can then instruct the AI models powering it to retrieve data.







