Apple developing photo verification tool for iOS 27

Apple is seemingly working on a photo verification tool that can help users determine the source of an image snapped by an iPhone camera.

Industry insiders unearthed references to the feature, called Apple Reference Image, in the iOS 27 beta 5.

As images generated by artificial intelligence become more widespread, the tool can help iOS users determine an image’s authenticity.

How it works

According to insiders, Apple Reference Image works by analysing a photo’s sensor information and metadata, or hidden text that describes the file itself, and sending it to Private Cloud Compute for verification. Then, the cloud network cross-verifies sensor data and assigns a picture a unique digital ID.

Aside from that, iOS users who receive authenticated photos in their Apple AirDrop or Messages can verify their validity using their own device.

Senders can also enable Transfer with Provenance when transferring photos from their iOS device to Mac or PC setups, enabling the digital ID to survive transfer from mobile to desktop.

The Apple Reference Image feature is said to be optional, with users asked to enable Reference mode for it to be available on their iPhone camera. 

Apple can also revoke a device’s authentication to maintain system integrity.

It is rumoured to arrive with the release of iOS 27 in September, but insiders said it may be rolled out later, around 2027.

READ MORE: Meta AI model goes rogue, hacks third-party company during security test

Image labelling in the age of AI

Apple’s move comes as the tech industry scrambles to label synthetic content, as AI-generated images flood the internet and fear about the spread of fake news and deepfakes persists.

In 2023, Google started developing SynthID, a watermarking tool that embeds digital watermarks directly in AI-generated media, including images, audio, text or video.

OpenAI, the parent company of ChatGPT, uses SynthID watermarks on top of C2PA metadata on files downloaded from its AI chatbot.

Meta started rolling out labels for AI-generated images in 2024, with the feature available on its social media platforms Facebook, Instagram and Threads.

In early 2026, X (formerly Twitter) began testing its “Made with AI” label for users to identify synthetic media.

READ NEXT: Vietnam’s VinSpace turns to SpaceX for maiden satellite mission