Meta is facing growing scrutiny after reports revealed that it is prompting Facebook users to enable a feature called “cloud processing,” which would allow the company to access and analyze private photos from users’ unpublished camera rolls.
According to a reports, this prompt appears when users try to post Stories on Facebook. It asks if they’d like to let Meta “regularly upload” their camera roll photos to generate AI-based suggestions, like collages, recaps, or themed content.
While Meta claims this feature is opt-in and not used to train AI models yet, it does not explicitly rule out future usage. Under the agreement, Meta can analyze facial features, timestamps, objects, and people in your photos — and retains the right to use that data.
Though Meta states that this feature is only analyzing the last 30 days of images, it admits that AI suggestions based on older media (like weddings or pet photos) may go beyond that limit.
Critics are comparing this to Google Photos, which makes AI suggestions but does not use personal photos to train generative AI. In contrast, Meta’s AI usage terms, updated in June 2024, leave a grey area about whether personal images uploaded via cloud processing could one day be included in training data.
Currently, users can turn off this feature and remove their uploaded media from Meta’s servers after 30 days. But the lack of transparency raises concerns about privacy and consent in the age of AI.
