Apple Photos can now add an invisible artificial intelligence watermark to pictures that users edit with certain Apple Intelligence tools. The change arrives with iOS 27.2, and it means a photo you took yourself can carry a hidden marker after an AI-assisted edit. The watermark is not visible on screen, and the person who receives a shared photo will likely notice nothing different.
Apple introduced the feature in its October 26 update. It relies on SynthID, a technology developed by Google, which embeds a digital signal into an image that compatible detection tools can recognize. According to Apple's official support documentation, the watermark is added when certain Apple Intelligence tools create or edit images. The marker does not change the appearance of the photo. There is no stamp across the bottom and no warning in the corner. The watermark stays hidden unless someone checks the image with a SynthID detector.
The affected tools include Clean Up, which lets users erase distractions from a picture while AI fills in the missing area. A family photo with a stranger in the background is the kind of example the feature's rollout calls to mind. Apple does not say that every small edit will trigger a watermark, and the marker can depend on the changes made to the image. Some photos edited this way can receive a SynthID watermark under iOS 27.2.
Apple's Extend tool is also covered. It generates additional scenery beyond the original edges of a photo, which can make a cropped shot look like a wider picture from the start. Because AI creates part of the finished image, the result can receive the hidden watermark. Reframe, which changes the apparent perspective of a picture after it is taken, is included as well. Image Playground and Image Wand, which create images from descriptions or rough sketches, are covered too. Since those tools generate visual content, their inclusion is straightforward. The harder question is why a genuine photo should receive an AI marker because one distracting object was removed.
Users who want to check whether a photo contains one of these markers can use an online detector. Apple recommends Google's SynthID Detector, which analyzes an uploaded image for the watermark. A detected watermark does not automatically mean a photo is fake. It shows only that AI was involved in creating or modifying the image. The reverse also holds: if no watermark appears, that does not prove the picture is genuine, since other AI image generators may not use SynthID and detection has its own limitations.
Privacy is a practical concern for anyone considering a check. Users should be careful about uploading private family pictures or sensitive images to any online service, and should review how that service handles uploaded files before submitting them. Apple says photos created or edited before the 27.2 update will not automatically receive a SynthID watermark, so the update will not add markers across an entire photo library. That changes, however, if an older picture is opened and edited with one of the covered tools after the update is installed.