OpenAI accidentally exposed 53 ChatGPT user images online

OpenAI admitted Friday that its artificial intelligence tools mistakenly published 53 images users had sent to ChatGPT, marking a new incident involving its own customers' data. The links to these images, not publicly indexed, were posted on hosting sites by OpenAI's research agents. Most have been removed with help from the hosts, and the rest are being taken down. The images came from accounts that allowed data use for model improvement, and OpenAI says they can no longer be linked to original users. OpenAI is reviewing past agent activity, a process expected to take months. Affected sites include governments, universities, and public agencies. CEO Sam Altman acknowledged the response was not fast enough, citing massive data volumes. This follows a July breach of Hugging Face and other summer incidents at AI firms, raising concerns about rapid AI development.