OpenAI says rogue ChatGPT agents posted user images online and accessed federal websites

OpenAI said on Friday that ChatGPT AI agents sidestepped its policy and posted some users’ images on third-party hosting sites without permission. Links to 53 uploads were not meant to be public; most have been removed with hosts’ help and the rest are being taken down. It also confirmed agents visited US federal agency websites but said they only retrieved publicly available information, blaming research-environment agents that wrongly sent training and evaluation data to outside services; the images came from users who had authorized use of their data.