Apple iPhone combats AI photo fakes with exciting camera feature in iOS 27

Even if a photo is taken with a camera and edited very little or not at all, it is becoming increasingly difficult for photographers to prove that it is actually a photo and not, for example, an AI-generated image. The Content Authenticity Initiative is tackling this problem with Content Credentials – when a photo taken with a camera that supports this feature is uploaded to the corresponding website, the site not only verifies the photo’s authenticity but also indicates how it was edited.

Apple is apparently working on a similar feature for the iPhone 17 ($759 on Amazon) and other models, as 9to5Mac discovered a feature called "Reference Image" in the fifth beta version of iOS 27. This feature is disabled by default and must be enabled in the System Settings under Camera before use. When the feature is active, after taking a photo in a special "Reference" mode in the Camera app, a RAW version, along with metadata and information about the user’s iPhone, is encrypted and uploaded to an iCloud server.

Apple itself reportedly does not have access to the photo but analyzes sensor metadata to ensure that the sensor has not been tampered with. If a verified photo is shared online or sent to friends, other Apple devices can determine whether the image is still verified or has been tampered with. This feature could be exciting not only for photojournalists but could also help combat scams, for example, when customers use AI to edit a photo of a new product to make it appear defective and defraud retailers out of a refund.