Joe Maring / Android Authority
TL;DR
- Google Messages appears to be working on a system to let users check if images in chat are AI creations.
- The system appears to distinguish between wholly AI imagery and cases where AI was only used for edits.
- None of this is yet functional, but Google’s already making progress.
Every day, we’re presented with an onslaught of images: browser ads, social media, news feeds — they’re everywhere. And increasingly, there’s a good chance that a plurality of those images might involve some kind of AI manipulation, if not full-on generation. While tools to recognize such pictures for what they are exist, oftentimes using them is far too manual of a process. But now it looks like Google could be streamlining how we access these solutions, especially when it comes to images sent through the Messages app.
Systems like SynthID let us scan media for AI content, and last year Google even baked SynthID into Gemini to make it easy to access across your device. Another of these is the Coalition for Content Provenance and Authenticity’s (C2PA) Content Credentials system, and Google Messages appears to be getting ready to analyze imagery for C2PA credentials.
Don’t want to miss the best from Android Authority?
Today we’re looking at version messages.android_20260611_04_RC00.phone.openbeta_dynamic of Google Messages for Android, but developers have been laying the groundwork here across a few recent releases. While we haven’t yet been able to see this C2PA-aware tool in action, text strings within the app hint at what’s coming:
Code
<string name="photos_threepio_infopanel_section_organization_text">Info by %s</string>
<string name="photos_threepio_infopanel_action_type_ai_title">Edited with AI tools</string>It sounds like we’ll find this accessible through the “View details” option in the overflow menu you’ll find after tapping on an image shared in a chat. Notably, it won’t just be an all-or-nothing “this is AI ” verdict, and the app is preparing to offer quite specific detail about the level of AI involvement. Here are some of the descriptions we’ve already found:
Code
Edited with multiple AI tools
Edited with multiple non-AI tools
May have been edited with multiple AI tools
Edited with non-AI tools
May have been edited with AI tools
Media made with AI
Media captured with a camera, multiple images were combined
Media captured with a camera without software adjustments
Media captured with a camera
Multiple pieces of media combined
Multiple pieces of non-AI media combined
Multiple pieces of media combined, some may have been made with AI
Multiple pieces of media made with AI combined
Multiple pieces of media combined including some made with AI
Media made with non-AI tools
Parts of this media were made with AI
Parts of this media may have been made with AI
Parts of this screenshot or recording may include AI contentIt might be even nicer if we could just see this kind of analysis right in out chat, rather than having to pull up image details — and who knows? Maybe Google could actually implement something just like that by the time this tool’s ready for a public release.
Even if it does mean a few taps, a quick way to learn about an image’s provenance sounds like a big upgrade for Messages in this messy AI-everywhere world we now inhabit.
⚠️ An APK teardown helps predict features that may arrive on a service in the future based on work-in-progress code. However, it is possible that such predicted features may not make it to a public release.
Thank you for being part of our community. Read our Comment Policy before posting.
