Gemini lets you verify AI-generated videos | Keryc
Google expands its transparency tools for AI-generated content and now you can check whether a video was created or edited with Google's tools directly from the Gemini app. That makes it easier to see when an audiovisual piece contains AI-generated elements without needing to be an expert.
What Google announces with Gemini
The change is simple and practical: you upload a video file and ask the app something like Was this generated using Google AI? Then Gemini looks for an imperceptible mark called SynthID in both the audio and the visual track.
The answer isn't a blunt yes or no. Gemini explains what it found and points to specific segments that contain elements generated with Google AI. For example, the system can indicate exactly that SynthID was detected in the audio between 10 and 20 seconds and that no mark was found in the visuals.
Maximum file size: 100 MB.
Maximum duration: 90 seconds.
Image and video verification available in all languages and countries supported by the Gemini app.
Example output showing the level of detail
SynthID detected in the audio between 10 and 20 seconds. No SynthID detected in the visuals.
How to use it in the Gemini app
Open the Gemini app on your phone or a compatible device.
Upload the video file (maximum 100 MB and 90 seconds).
Ask a clear question, for example: Was this generated using Google AI?
Check the answer: Gemini will tell you if it detected SynthID and in which segments of the audio or visual track.
It's an experience designed to be direct: you don't need technical knowledge to get a clue about whether a video contains elements generated by Google's tools.
Why this matters and what its limits are
Why should you care? Because we live in a world where deepfakes and AI edits are increasingly accessible. Having a tool that indicates the presence of an origin mark helps journalists, creators, and everyday users evaluate the authenticity of material.
But heads up: this isn't a silver bullet. Some important considerations:
SynthID identifies content generated or edited with Google technologies that embed that mark. It doesn't detect every form of AI or marks from other companies.
The absence of a mark doesn't guarantee the content is 100 percent human. The mark might not have been present or it could have been removed.
The 100 MB and 90-second limits exclude long or high-resolution videos; for those you'll need alternative verification methods.
In short, it's a useful tool to add context and clues, not an infallible proof on its own.
A pragmatic step toward more transparency
The arrival of verification inside the Gemini app is a clear sign: big platforms are integrating ways for the public to better understand when AI has intervened. Do you share videos, work with digital content, or just want to check what you receive? This feature gives you a quick and accessible way to get context.
