Skip to content
Tech News
← Back to articles

Google Messages just made it easier to spot AI-edited photos in your group chats

read original more articles
Why This Matters

Google Messages' new Content Credentials feature enhances transparency by providing a digital history of media files, helping users identify AI edits and modifications in shared photos, videos, and audio. While it doesn't verify authenticity, this tool offers valuable context in an era of increasingly sophisticated AI-generated content, aiding consumers and the industry in detecting potential misinformation. This development underscores the importance of digital provenance in maintaining trust and integrity in digital communications.

Key Takeaways

Sanuj Bhatia / Android Authority

TL;DR Google Messages now supports Content Credentials, letting you check how compatible photos, videos, and audio were created or edited, including whether AI tools were involved.

It won’t tell you whether something is real or fake. Content Credentials only show the history attached to a file, and media without credentials won’t provide this information.

The feature is currently Android-only. Press and hold supported media, tap Info, and look for How this was made to see its origin and editing history.

A couple of months ago, we spotted Google Messages preparing a useful new trick for the age of AI: a way to tell you about where the photos, videos, and audio people send you actually came from. That feature is now available, and it could come in handy the next time something in your group chat looks too good to be true.

Google Messages now supports Content Credentials, which essentially give compatible media a digital history. Instead of leaving you to stare at a suspicious photo and wonder whether it was shot with a camera, heavily edited, or created with the help of AI, Messages can show you some of that information directly.

Joe Maring / Android Authority

These credentials are based on the C2PA standard. The easiest way to think about it is as a digital paper trail attached to a photo, video, or audio file. So, imagine a friend sends you a beautiful sunset photo. Content Credentials could tell you that the image started life as a camera photo but was later edited using AI tools . In some cases, Messages may know that an image was edited but won’t have enough information to say whether AI was involved, in which case it’ll tell you that it “may have been edited by AI tools.”

There is one important limitation, though. Google Messages can’t tell you whether something is real or fake. It can only show you the history recorded with that file. So, if the file contains Content Credentials showing that AI tools were used to edit it, you now have a pretty useful clue before forwarding it to someone else.

The feature also won’t work with everything you receive. If that information was never attached to the file, Google Messages won’t have anything to show you. So, importantly, the absence of Content Credentials doesn’t mean a file is authentic or hasn’t been edited.

... continue reading