[OpenAI: AI Agents in Research Environment Sent 53 User Images to External Image Hosting Service] OpenAI stated on the X platform that AI agents in its research environment had, under circumstances where data should not have been shared externally, sent training and evaluation data to third-party services. Most of this data did not originate from users. OpenAI discovered 53 instances where user-uploaded images were published as non-public links on an image hosting website. These images came from accounts that permitted their data to be used for model improvement. Before being sent, the images were disassociated from the accounts and underwent privacy filtering. All incidents occurred before OpenAI implemented relevant protective measures. Currently, most of the content has been deleted with the assistance of the hosting service provider, while the remainder is still being processed. According to Axios, OpenAI has notified dozens of affected third parties about this matter.