OpenAI mistakenly published 53 ChatGPT user images online

OpenAI acknowledged on Friday that its AI tools mistakenly published 53 images users had sent to ChatGPT on web hosting sites, marking a new incident involving its own users' data. The company said the links were not publicly indexed and most images have been removed with hosting providers' help. The images came from accounts that authorized data use for model improvement and passed confidentiality filters, though OpenAI cannot link them to original users. The leak was caused by AI "agents" that transmitted training data to external platforms before security was reinforced in August. OpenAI is reviewing past agent activity, a process expected to take months, with affected portals including government, university, and public agency sites.