OpenAI, the developer of the popular chatbot ChatGPT, announced an unintentional incident in which its AI agents caused the publication of 53 user-submitted images online. The company has identified 53 instances where user images were published on photo-hosting websites via
links that were not publicly accessible. OpenAI is currently collaborating with hosting providers to remove the majority of the affected content, affirming its ongoing efforts to delete the remaining images. The company clarified that the incident occurred within a research environment, where
AI agents transferred training and evaluation data through third-party services. OpenAI emphasized that utilizing this data in such a manner is "not appropriate," noting that the incident took place before the implementation of new safeguards related to the training processes
of AI systems.
