Making it easier to understand how content was created and edited

Summary

Google is expanding content provenance and verification capabilities across Search, Gemini, Chrome, Pixel and Google Cloud. Key technologies include SynthID digital watermarks and C2PA Content Credentials for capture/edit provenance; SynthID verification is currently in the Gemini app and will roll out to Search and Chrome. Google is also launching an AI Content Detection API on the Gemini Enterprise Agent Platform for customers to detect AI-generated media from Google and third-party models.

Key Points

• SynthID watermarking has been integrated into Google’s generative media stack and applied at scale (over 100 billion images/videos and 60,000 years of audio). Verification is available in the Gemini app (used 50M times) and is expanding to Search and Chrome.
• C2PA Content Credentials record when media was captured or modified. Pixel 10 already embeds image credentials; video Content Credentials will be added to Pixel 8, 9 and 10 soon. C2PA verification is rolling out in Gemini and coming to Search/Chrome.
• Google is launching an AI Content Detection API on Google Cloud’s Gemini Enterprise Agent Platform to detect AI-made media from Google and other popular models. Target use cases: feed sorting, fraud prevention, fact-checking, and user-facing labeling.
• Industry adoption and standards: partners (OpenAI, Kakao, ElevenLabs, NVIDIA) are adopting SynthID and Google has open-sourced text watermarking. Google continues to work in the C2PA steering committee to ensure interoperability (e.g., Meta will label camera-captured media with Content Credentials on Instagram).

Engineer notes / recommended actions

• Evaluate integrating the AI Content Detection API into ingestion and moderation pipelines to tag or route suspected AI-generated media.
• Update metadata ingestion to consume SynthID signals and C2PA Content Credentials where available; preserve provenance fields through processing and storage.
• Add verification UI/automation for users or downstream systems to query “is this made with AI?” via Search, Gemini, or Chrome endpoints as they become available.
• Monitor rollouts (Gemini app now, Search/Chrome over coming weeks/months) and Pixel device updates for capture-side metadata on images and video.

Link

Original announcement published 2026-05-19.