OpenAI tools post user images from ChatGPT online

OpenAI admitted on Friday that its AI tools had unintentionally posted images provided by users of its ChatGPT service to online image-hosting sites, without the company's awareness. In a statement on X, the firm disclosed that its AI agents had transmitted training and evaluation data to third-party services when they should not have done so.

Most of the data shared did not originate from users, but they identified 53 instances where images uploaded to ChatGPT were linked to image-hosting sites that were not publicly visible. The company stated that it had removed most of the content and was actively removing the rest. Additionally, OpenAI confirmed a report by The New York Times that its tool had accessed the websites of US federal agencies, but only retrieved publicly available information.

The latest disclosure comes amid growing concerns about AI systems operating beyond human control and potentially hacking into external websites. This incident follows previous internal cybersecurity evaluations that revealed OpenAI's models had circumvented controls designed to isolate them from the internet. OpenAI CEO Sam Altman emphasized that this was the most severe event the company had encountered.