Google is developing an AI image detection capability for Messages that enables users to verify if photos exchanged during chats originated from or were modified by AI tools, reports Android Authority. Although the capability remains inactive, analysis of code in a recent beta version reveals plans for C2PA Content Credentials support.

More than a simple AI label

C2PA provides an established framework that tracks the creation and alteration history of media files. Content Credentials embed metadata within images to document the originator, applications involved, and successive changes performed. Incorporating this system into Google Messages would offer recipients a clear method to evaluate the source of received visuals. Notably, the approach avoids limiting results to a basic AI versus non-AI classification.

Discovered code references include specific indicators such as Media made with AI, Edited with multiple AI tools, Parts of this media were made with AI, and Media captured with a camera without software adjustments. These granular descriptions would enhance the feature value compared with generic alerts.

Still in development

Google already incorporates its SynthID watermarking system within Gemini for AI content identification on Android devices. In contrast to that proprietary method, C2PA represents an open standard supported by organizations including Adobe, Microsoft, and Qualcomm, allowing broader detection across varied sources.

Android Authority indicates the option will probably appear via a View details selection within the menu accessed by tapping a shared image. All information stems from an APK examination, so the capability is not yet operational and could evolve prior to release. No official confirmation has been issued by Google, leaving the timing for wider availability uncertain.