While everyone else is busy slapping labels on AI-generated images, Apple appears to be building the reverse system.
According to 9to5Mac's analysis of the latest iOS 27 beta, traces of a feature called Apple Reference Image have surfaced in the system. What it's trying to verify isn't "is this image AI-generated" — it's "was this photo actually captured with an iPhone camera." To get this certification, users would first need to switch on a new Reference mode in the Camera app; after shooting, the system would send partial sensor data and metadata to Apple's Private Cloud Compute to confirm the image's origin.
This approach runs completely counter to where the industry is currently headed. Apple's own Image Playground and other upcoming photo-editing features follow the digital watermark route — stamping AI-generated content with a "this is synthetic" mark, similar to what Google's SynthID initiative has been doing and continuing to expand. Reference Image flips the script: rather than flagging what's fake, it gives genuinely captured photos a verifiable identity. If both approaches end up running side by side, they could theoretically complement each other — one system tagging what's artificial, the other proving what's real.
Everything known so far comes from a beta build still in testing, and 9to5Mac's findings are inferences drawn from system signals about how this feature might exist and function — Apple has not officially confirmed any of it. That means Reference Image could still be absent when iOS 27 officially launches this fall, or it might show up in a completely different form.
