iPhone and AI-generated images: Apple is working on a badge to certify authentic photos

Images made by artificial intelligence are flooding social networks, and the industry is organizing itself to label them: invisible watermarks, labels, provenance metadata. Apple is exploring the opposite path, that of certifying the genuine rather than tracking the fake. A confidentiality disclosure slipped into the fifth beta of iOS 27 reveals a system called “Apple Reference Image”, capable of certifying that a photo was indeed taken with an iPhone.The feature is not yet active as of today.

Also read: The all-glass iPhone of the future may not be dead, but Apple’s project would have changed

A badge that traces back to your iPhone’s sensor

The principle relies on a “Reference” mode that must be manually enabled in the camera app. Photos captured in this mode embed provenance data directly into their metadata, like a jeweler’s hallmark stamped at the time of capture. To verify an image, all you need to do is tap its badge: the raw image, sensor signatures, capture window and hardware identifiers are then sent to the Private Cloud Compute servers. These servers determine whether a physical sensor actually produced the photo, assign it a unique identifier and return an authenticated version.

Apple says it only receives the sensor data, a fingerprint and the verdict along the way, never the photo itself. Above all, the company reserves itself an unprecedented enforcement power: to refuse to authenticate images from a sensor deemed compromised, and even to retroactively revoke certifications already granted to a device. The intended audience is easy to guess: professional photographers, agencies and newsrooms, especially since the text also mentions “uncropped sequences,” suggesting a future video section. No date accompanies any of this, and a feature identified in beta can still change form or even disappear before the final version.

Everyone marks the fake, Apple prefers to stamp the real

This approach runs counter to everything the industry has deployed over the past two years. Image generators label their creations (those from Image Playground already carry their marking in the metadata), and, since August 2, Europe has required AI-generated content to be identifiable by machines. Certifying the genuine capture completes this system at the other end: on the one hand, seals are placed on the fake; on the other, a certificate of origin is issued, much like a controlled designation for photography.

Read also: How to know whether an image is authentic? Google has a solution

The lone rider’s choice, on the other hand, will come as no surprise. Leica, Sony and Nikon embed their proof of authenticity in the common C2PA standard, which Google has joined with its Pixel 10. Apple, meanwhile, remains absent from the consortium and is building an equivalent but entirely in-house system, verified by its own servers (the company has never been fond of standards it did not invent). Nothing in the beta mentions the common standard, either. The system will also have its limitations: the badge certifies what the sensor saw, not what the photo became afterward. A screen displaying a fake, photographed in Reference mode, will indeed yield a perfectly authentic image… of a counterfeit.

👉🏻 Follow the tech news in real time: add 01net to your feeds on Google, and subscribe to our WhatsApp channel.

Source:Mac Rumors