Facebook users are being urged to exercise caution before enabling a new feature that allows Meta, Facebook's parent company, to access and scan photos stored on their phones — including those never shared on social media platforms.
The development follows growing concerns over Meta’s use of user data, especially after reports confirmed that the company has been training its artificial intelligence (AI) models using publicly shared photos from Facebook and Instagram.
However, recent revelations indicate that Meta now seeks access to private photos stored on users’ devices, according to a report by asianetnews.
A TechCrunch report, cited by India Today and The Verge, explains that some Facebook users recently received pop-up notifications while attempting to upload a story. The notification offered the option to activate a new feature called Cloud Processing, which enables Meta to automatically upload photos from a user’s camera roll to the company’s cloud.
The feature promises to offer users AI-powered creative tools such as photo collages, event recaps, AI-generated filters, and theme-based suggestions for occasions like birthdays or graduations.
While the feature may appear useful and harmless at first glance, experts warn of significant privacy risks. Once activated, users effectively give Meta permission to scan, analyze, and process personal photos stored on their devices, including those never posted online. Meta’s AI system will reportedly examine faces, objects, locations, dates, and even the metadata embedded in those images.
Meta has defended the feature, describing it as an entirely optional service aimed at enhancing user experience. The company says users can turn the feature on or off at any time. “It is an opt-in feature that you can turn on or off at will,” Meta said.
Rise in harmful content on Facebook following Meta's moderation rollback
Despite these assurances, privacy advocates remain concerned, especially considering Meta's history of handling user data. The company recently admitted that it has been using all photos shared publicly on Facebook and Instagram since 2007 to train its generative AI models.
However, Meta has not clearly defined what qualifies as ‘public’ content or what age restrictions apply to its data use policies, raising further questions.
To opt out of Cloud Processing, users can disable the feature through Facebook’s settings. Meta says that if the feature is turned off, any unshared photos uploaded to the cloud will be deleted within 30 days.
As tech companies continue to experiment with the limits of user data collection in the AI era, experts warn that features like Cloud Processing — though presented as tools for user convenience — may quietly expand access to personal data.
Previously, users had to consciously decide to share photos publicly. With Cloud Processing enabled, however, those same photos can be silently uploaded to Meta’s servers, allowing Meta AI to access them.
In this context, experts advise users to carefully review the terms of such features and make informed decisions to protect their privacy.