Users who want to share a post on Facebook are greeted with a new warning: a pop-up window asking for permission for “cloud-based processing.” If we approve, the system can access our entire phone photo library—including photos we've never uploaded to the social network. The goal: to generate creative ideas using artificial intelligence, such as collages, themed selections, or stylized versions.
At first glance, the feature may seem harmless. Users receive suggestions in their own gallery, which they only share with others if they consciously choose to do so. Meta emphasizes that these suggestions are made exclusively for the individual and are not used for advertising purposes. The feature can also be disabled and is only activated if someone explicitly agrees to its use.
Still, it is worth dwelling on the details in several respects. By pressing the “allow” button, the user accepts Meta's terms of service for artificial intelligence. According to these terms, the company processes not only the images themselves for analysis, but also their metadata, such as the time and location they were taken, and the people and objects they depict. The system may also use facial recognition technology and may not only use the data for the functioning of the feature, but also store it for longer periods of time.
According to Meta's statement, this data is not currently used to train AI models. However, they did not give a clear answer to the question of whether this would be ruled out in the future. This uncertainty is not without precedent. In 2024, the company acknowledged that it had widely used publicly available content—posts, images, comments—published over the past decade and a half to develop its generative artificial intelligence. Exactly what they considered “public” and who counted as an “adult user” from the beginning remains unclear to this day.
The current development goes beyond previous practice in that it requests access not only to published photos, but also to private ones. European Union regulations required Meta to provide an option to opt out of AI training, but this was only available until May 27, 2025. Those who did not opt out by then are now likely to have their photos included in the analysis if they are uploaded to the cloud.
Meta sees the initiative as a test. They are trying out automatic suggestions among users in the United States and Canada, which they claim are offered solely to simplify content sharing. Detailed help documentation is available for the service, and most settings are indeed customizable.
However, the introduction of the feature has been met with little public debate. Most users encounter it when they first see the pop-up window, and many do not understand exactly what approval means. Examples have already appeared on some forums of Facebook automatically converting old, previously uploaded images into anime style, or making suggestions that were surprising even to the user.
The use of social media and the development of artificial intelligence undoubtedly offer new opportunities. At the same time, it is worth taking a critical view of solutions that take the handling of personal data to a whole new level. Our photo albums are not just for storing pictures—they contain faces, places, relationships, and memories. A system that automatically processes and transforms them, even if it appears to serve creativity, requires careful consideration before we entrust it with the visual imprints of our private lives. Perhaps the best or most conscious user response to the emergence of such features is to immediately delete your Facebook account.
Sources:
