Google is making it easier to identify AI-edited images by adding invisible digital watermarks to photos modified with its Reimagine tool.
What is Reimagine?
Reimagine is part of Google Photos’ Magic Editor, introduced with the Pixel 9 series. It allows users to change parts of an image using a simple text prompt. While these edits can be fun and creative, they can also be misleading.
How Does the AI Watermark Work?
Google is now embedding an invisible watermark called SynthID into images edited with Reimagine. This technology, developed by Google DeepMind, helps users recognize AI-generated changes by checking the “About this image” section in Google Photos.
For example, AI-edited images will display an “AI info” section stating:
✔ Credit: Edited with Google AI
✔ Digital source type: Edited using Generative AI
Users can also verify images outside Google Photos by using Circle to Search, which detects AI-generated elements.
Why is This Important?
These hidden watermarks make it harder to spread fake images on social media and messaging apps. Unlike regular tags, SynthID watermarks are embedded into the image itself, making them difficult to remove.
Limitations of SynthID Watermarks
While the system is designed to be durable, repeated edits may weaken the watermark. Small changes, like altering a tiny part of an image, might go undetected.
Google already uses SynthID for AI-generated images and is now applying it to all pictures edited with Reimagine. Currently, Reimagine is available only on Pixel 9 and newer devices, but all devices can detect SynthID watermarks.
This move is part of Google’s effort to make AI-generated content more transparent and reduce misinformation online.
Source: forbes
