Google Built a Tool to check if a Video was Generated on a Google Platform

Lately, there has been a growing difficulty in identifying whether a video is AI. That’s why Google has added something called a SynthID watermark to all its AI-generated content, which is a digital tag embedded in both audio and video tracks. That’s what a new feature from Google in the Gemini app uses to check if a video was generated by a Google AI model.

The feature lets you upload a file up to 100 MB in size or 90 seconds long, then you can simply ask it, “Was the file generated using Google AI?” Thankfully, Gemini doesn't just give a Yes or No; it also breaks the video into segments. It could also see that the visuals are real, but the audio was AI-generated between specific timestamps, which is cool that it can do that.

This feature is actually not in Beta; it will be available in all languages and countries where the Gemini app is available, making AI video verification more accessible to everyone.

Could we actually get to a world where this would be used for almost all the videos we watch? Will this feature be useful to you? Have you ever been fooled by an AI-generated image or video? If so, what was the image or video? Let us know in the comments!

Source: Google