iPhone AI images and AI photo edits now carry a hidden Google watermark, Apple says

Pictures that iPhone owners create or touch up with Apple's AI tools now carry a hidden watermark that marks them as AI-made, according to an Apple support document published Nov. 5.  The change comes with iOS 27.2, the current iPhone software, and the matching 27.2 updates for iPads, Macs and the Vision Pro headset.

Which pictures get the mark

The watermark is added by five features in Apple Intelligence, Apple's set of AI tools.  Image Playground and Image Wand make new pictures.  Clean Up, Extend and Spatial Reframing are AI editing tools in the Photos app.  That means a real family photo gets the mark too if one of those tools changes it.  Older pictures are not marked unless they are edited again with one of those tools, Apple says.

How it works

The mark comes from SynthID, a technology made by Google DeepMind, Google's AI lab.  It hides a signal inside the picture itself that people cannot see and that does not change how the picture looks, but that special detection software can read.  That is different from the visible "AI" labels some apps add, which can be cropped off.  To check a picture, Apple tells users to upload it at synthid.com.  Apple's page does not describe a way to check inside the iPhone itself.

Why it matters

As AI-edited photos get harder to tell from real ones, tech companies are adding marks like this so news outlets, social media sites and ordinary people can check what a picture is.  Google says no watermark is foolproof, and its own documentation says SynthID is not built to stop a determined bad actor, only to make misuse harder.  To get the new mark, iPhone owners need to update to iOS 27.2 in the Settings app.

Source:  Apple support page on SynthID watermarks; Apple software versions; Google's SynthID documentation