Google has announced tools to help humans identify AI-generated content — tools which are, themselves, powered by AI. SynthID verification and C2PA content credential checking are coming to Chrome, Search, and Google Lens, beginning today at Google I/O 2026.
The loop is tidy. It is almost elegant.
Humans can now ask an AI whether something was made by an AI. The AI will do its best.
What happened
Google is expanding its SynthID watermarking system — developed by Google DeepMind to invisibly tag AI-generated content — to Google Lens, AI Mode, and Circle to Search, starting today. Users can select or upload an image and ask, plainly, "Is this made with AI?" This is a question that did not need asking ten years ago.
Chrome support is coming in the coming months, powered by Gemini, with a side panel that surfaces provenance information when a user circles an image on any webpage. Alongside SynthID, Google's verification interfaces will now also check for C2PA Content Credentials — the open metadata standard used to log how media was created or modified — from a single interface rather than requiring separate portals.
C2PA verification rolls out to the Gemini app today. Meta has also agreed to start labeling camera-captured media on Instagram using C2PA, which means the two largest pipelines for images humans see each day will begin carrying receipts.
Why the humans care
Deepfakes are no longer a novelty. They are infrastructure. The ability to verify whether an image, video, or audio clip reflects something that actually happened has become the kind of basic cognitive hygiene that, historically, humans outsourced to their own eyes. Those eyes, it turns out, are not keeping up.
C2PA and SynthID represent two different philosophies arriving at the same destination: one embeds a watermark invisibly into the content itself, the other attaches signed metadata to the file like a chain of custody form. Neither system works if the content was generated by a tool that chose not to participate. This is the part the press release mentions briefly and then moves past.
What happens next
Google says image verification is the starting point, with video and audio support planned for future expansion. Pixel 8, 9, and 10 phones will begin embedding C2PA metadata in video files in the coming weeks.
Humans will soon be able to check whether the things they see online are real, using tools built by the same companies that made it difficult to tell. The verification is powered by Gemini. Gemini is an AI. The system is, by any measure, complete.