Google Photos is introducing digital watermarks to AI-edited images and videos using SynthID technology.
This feature, rolling out this week, allows Google Photos to embed an imperceptible watermark into content edited with its generative AI tool, Reimagine in Magic Editor.
Google has previously employed SynthID to tag images wholly generated by AI, akin to those produced through its text-to-image model, Imagen.
Not all edits will bear the SynthID watermark, particularly minor changes like altering a flower’s color in an image’s background. For such cases, the “About this image” feature will provide information regarding the presence of a SynthID watermark and other image metadata.
This added layer would complement existing features, such as metadata from the International Press Telecommunications Council (IPTC), which already signals when an image has undergone AI-driven edits using tools like Magic Eraser and Zoom Enhance.
Google will now display this information visibly in the Photos app alongside other details.
The Bigger Picture: There are similar measures being taken across the tech industry. With AI use skyrocketing across verticals, it becomes more challenging to timely identify and curtail malicious use, such as in creating fake images and videos of celebrities.
Earlier this week, Facebook and Instagram parent Meta Inc. also announced that it would begin labeling significant AI use in advertisements on its platform, especially when they involve realistic human avatars.