Apple is reportedly developing an iOS feature that would let iPhone users prove their photos are genuine and not AI-generated fakes. Code references for an "Apple Reference Image" system have been spotted in the iOS 27 beta 5, which would embed provenance metadata into photos at the moment they are captured, allowing their origin to be authenticated later. The move would put Apple in line with a growing industry push to distinguish real images from deepfakes, though it has chosen to build its own system rather than adopt the existing C2PA standard used by rivals.

The feature is not yet live and would be switched off by default, reportedly accessible via Settings > Camera > Reference Image once enabled, and only photos taken with a new "Reference" camera mode would carry the necessary data. Authentication would not be automatic: users would tap a "Reference" badge, prompting the image and metadata—including sensor signatures, capture time and hardware identifiers—to be sent to Apple's Private Cloud Compute servers, which return an authenticated version with a unique ID. Apple says it would not access the raw photo itself, though it could use sensor data to block or revoke authentication for compromised devices. This differs from the C2PA Content Credentials standard already used by Canon, Nikon, Sony, Fujifilm, Leica and Google's Pixel 10, which Apple has so far avoided adopting.