Apple's Reference Image tool lets you check if an iPhone photo was edited or AI-altered

Apple has quietly rolled out a new feature called Reference Image that aims to answer a question growing more urgent by the month: was this photo edited, and was AI involved? The tool, spotted in Apple's latest software, gives users a way to inspect an image for signs of alteration, including changes made by generative AI. It arrives as fake and manipulated photos spread across social platforms faster than platforms can label them.

The mechanics matter less than the intent. Apple is building the check into its own ecosystem rather than leaving it to third-party apps or fact-checking sites. That puts the company in the position of arbiter for its own camera output, a role it has mostly avoided. The feature likely leans on metadata and image forensics rather than a cloud lookup, though Apple has not detailed the exact method.

The timing is not accidental. Google, Adobe and Meta have all shipped content credentials or invisible watermarking schemes over the past two years, and camera makers like Leica and Sony have added similar signing tech. Apple's entry raises the stakes because the iPhone is the most-used camera in the world. If Reference Image works reliably, it could become the default way hundreds of millions of people check a photo before sharing it.

There is a catch. Any verification tool is only as good as the edits it can detect. Cropping, color grading and screenshotting can strip metadata, and AI models improve at erasing their own traces. Apple has not said whether Reference Image will flag edits made by non-Apple apps, or whether it can distinguish a light retouch from a full generative replacement. Those gaps will decide whether the feature is a real safeguard or a comfort blanket.

For now, the feature is another sign that provenance is becoming a platform-level concern rather than a niche research problem. Apple has not announced a broader content-credentials standard, and it has not said whether Reference Image will feed into Messages, Photos sharing or third-party apps. The open question is whether users will trust a verdict that comes from the same company that made the camera.