"Here Come the AI Worms"

A team of researchers developed one of the first generative Artificial Intelligence (AI) worms, which can spread from one system to another. Ben Nassi, a Cornell Tech researcher, together with fellow researchers Stav Cohen and Ron Bitton, developed the AI worm dubbed "Morris II." They demonstrated how the AI worm can attack a generative AI email assistant, stealing data from emails and sending spam messages. The team was able to break some security protections in ChatGPT and Gemini. The study, which was conducted in test environments rather than against a publicly available email assistant, comes at a time when Large Language Models (LLMs) are becoming more multimodal. Although generative AI worms have not yet been observed in the wild, researchers say that startups, developers, and technology companies should be mindful of this security risk. This article continues to discuss the study "ComPromptMized: Unleashing Zero-click Worms that Target GenAI-Powered Applications."

Wired reports "Here Come the AI Worms"

Submitted by grigby1