Facebook is testing a feature that allows AI to process users’ photos before they are published
Facebook is testing a new feature that allows Meta artificial intelligence to analyze users’ photos before they are uploaded to the social network. It is noted that when trying to publish a new story on the platform, a message may appear asking for permission to create collages, stories or photo redesigns based on images from the user’s gallery. About this informs TechCrunch.
By giving consent, the user accepts Meta’s terms, which allow its AI to analyze not only published, but also unpublished photos — including facial recognition, the date of shooting, the presence of other people or objects in the frame. In addition, the company gets the right to store this data and use it in the future.
Despite Meta’s statement that this data will not be used for advertising purposes, there is a possibility that the images could be used to train artificial intelligence models. The company denies this. In a comment to The Verge, Meta representatives said that no photos are currently being used to train AI models. According to them, the camera footage can only be used to improve personalized features, but not for training.
Meta emphasizes that testing of this feature is limited, although mentions of it appeared as early as the beginning of the year. For users who previously agreed to the processing of photos from their gallery, the option to disable this function is available through the “Help Center”.
