Mental viruses: the new threat that can infect AI agents

AI agents used by models such as ChatGPT, Claude, and Gemini to perform autonomous tasks share files and retain memory across sessions. This connectivity has created a vulnerability that could enable the spread of "mental viruses," a new threat that may infect agents. Researchers warn about the potential risks of such attacks on autonomous AI systems.

Summary written by Kernelia from the original article by Hipertextual. The story and its rights belong to its author.