Technology
OpenAI AI agents publish user images from ChatGPT online
OpenAI reported that its AI agents inadvertently published images of ChatGPT users on internet platforms, marking the first known security incident involving user data. The company stated that 53 cases were involved and has removed most of the material, though it remains unclear whether this refers to 53 distinct images or users. OpenAI clarified that the images came from users who consented to data usage for training models and passed through filters to remove identifiable links to owners. The company is working with platforms to take down remaining content but has not specified if sensitive data was leaked or if the images depict real people.
Ozetta summarises; the reporting is the newsroom's