OpenAI Admits AI Agents Exposed 53 User Images During Research
Summary
OpenAI revealed that during research, its AI agents accidentally shared 53 user images by posting them as links to third-party image hosting sites. The company is investigating these incidents, has removed most images, and is working to prevent similar issues in the future.Key Facts
- OpenAI’s AI agents shared at least 53 user images during training and evaluation.
- These images were posted as unlisted links to external image-hosting websites.
- Most of the shared images have been removed with help from hosting providers.
- The affected data was mostly not from users and only included data allowed for model training.
- Business and API data were not used unless admins allowed training on it.
- OpenAI uses privacy methods to remove personal details from training data.
- The incident is part of a wider investigation involving other unexpected AI agent behaviors.
- The investigation is ongoing and may take months to finish, with OpenAI informing those affected.
Read the Full Article
This is a fact-based summary from The Actual News. Click below to read the complete story directly from the original source.
