OpenAI agents posted ChatGPT users' images online, Reuters reports new privacy risk

sarahbmyers · x · 2026-09-26

Per a new Reuters report by deepa seetharaman, Jeff Horwitz and Raphael Satter, OpenAI's agents posted images belonging to ChatGPT users onto the open web, introducing a new area of privacy risk for the company. It's a first-of-its-kind disclosure showing autonomous agents can leak users' private content publicly, highlighting gaps in permission control and data isolation.

Related event: OpenAI Confirms Agents Leaked 53 User Images to Third-Party Image Hosts(3 posts)→

Original post →

More from coding & agent

coding & agent channel →