Apple is building a system called Reference Image to verify that photos actually came from an iPhone camera rather than being AI-generated or heavily manipulated. The feature works by tying authentication to unique hardware identifiers from the iPhone's camera sensor itself. When you take a photo using a new "Reference" mode in the Camera app, it captures sensor data and metadata that can later be verified through Apple's Private Cloud Compute servers. The whole process is designed so Apple never sees the actual photo—only the sensor information and metadata get sent for verification. The authentication is opt-in and off by default. You have to enable it in Settings under Camera > Reference Image, and crucially, you need to shoot photos specifically in Reference mode for them to be authenticatable later. Regular photos won't have the provenance data needed. When you later choose to authenticate an image, Apple assigns it a unique ID and can confirm it came from a legitimate camera sensor. If Apple discovers a sensor has been compromised, it can revoke all previous authentications tied to that hardware. Where this gets interesting for sharing: authenticated photos can be verified on other people's Apple devices locally, without Apple knowing which images are being checked. You can share via AirDrop, Messages, or USB transfer (with a new "Transfer with Provenance" option for Mac and PC). The system tracks whether you're sharing the original, an edited version, or just the authentication status. If someone gets the full photo data package, it includes hardware identifiers and uncropped footage. This is Apple's answer to the AI-generation problem that's already prompted responses from Google (SynthID watermarking) and Meta (content labels). The difference here is Apple's focus on hardware-level authentication rather than watermarking or post-hoc labeling. The code for Reference Image appeared in iOS 27 beta 5, so the feature isn't live yet, but the groundwork is clearly in place.