Alex Omenye

Facebook is testing a new feature that encourages users to grant the app access to their phone’s entire camera roll, including photos never uploaded to the platform, in order to generate AI-edited content.

The prompt appears when users attempt to create a new Story, asking them to opt into “cloud processing” that enables the app to offer personalized creative suggestions.

By agreeing, users allow Facebook to upload photos from their device to Meta’s servers on an ongoing basis. The AI system then uses the data—such as timestamps, locations, and visual themes—to generate collages, recap posts, restyled images, or themed suggestions. While Facebook assures users that only they can see these outputs and that the data isn’t used for ad targeting, the underlying permissions raise broader privacy questions.

Accepting the feature also binds users to Meta’s AI Terms, which permit the company to analyze image content, including facial features, and retain the data to personalize future AI interactions. Meta also reserves the right to review AI user interactions—potentially including photos, prompts, and other shared content—with human oversight.

Though the feature has drawn little widespread backlash so far, some users have expressed concern on Reddit and Facebook groups. One reported that Meta automatically transformed an old photo into an anime image without explicit instruction.

The feature can be toggled off in the app’s settings under “Camera roll sharing suggestions.” Two options appear: one for photo suggestions and another for cloud processing.

Meta did not respond to requests for comment. The tool signals a broader shift in how personal media may fuel AI development, raising questions about consent and transparency, especially as many users may not fully understand what they’re agreeing to in the rapid rollout of such features.