Google is expanding its content transparency tools within the Gemini app. It is now possible to verify images generated by the company's proprietary artificial intelligence model. Note that this is limited to those created by Gemini, as the AI uses an ID that only you can see.
It's becoming increasingly difficult to tell whether the video you receive is a real recording or one cooked by a computer, but now you can upload it and check it. The process is very simple. Simply upload your video to Gemini and ask a simple question like “Was this generated using Google AI?”
Gemini will then start scanning something called SynthID. This is Google's proprietary digital watermarking technology. Embed signals in AI-generated content. This signal is imperceptible to humans, but can be easily detected by software. This tool thoroughly checks the entire file to see if AI is used in the background music, the footage itself, or both.
The response you get is not just a simple “yes” or “no.” I think Gemini is helpful because it uses its own reasoning to provide context. You also specify which segments of your content include the AI-generated elements. For example, you might see a response that says “SynthID detected in 10-20 seconds of audio. No SynthID detected in visual.” This level of detail is a great feature for anyone trying to understand what's true and what's not in the media.
If you use this tool frequently, there are some practical limitations you should be aware of. Currently, uploaded files cannot exceed 100 MB in size and execution time cannot exceed 90 seconds. This means you won't be checking out full-length movies, but it's perfectly sufficient to check out short clips and social media content.
This new video verification feature expands on Google's previously launched tools for images. The company has been promoting its SynthID technology for some time, aiming to establish transparency in the content generated by its tools. Since its introduction in 2023, the tech giant claims it has watermarked more than 20 billion pieces of AI-generated content. This large amount of marked content means that Gemini can almost instantly identify an AI image if it comes from a Google generator. This expansion brings the same level of scrutiny to movies and audio.
However, I must speak of a major caveat. This tool is strictly limited to content generated or edited using Google's own internal tools. If an image or video was created using an AI model operated by someone other than Google, Gemini can't tell you anything about it. This means that this tool is really only useful for ensuring transparency within Google's own ecosystem.
Rather than running your images and videos through a third-party checker, Google wants you to rely on Gemini for these checks. Although the company makes it easy to identify its own work, it should not be considered a general-purpose AI detection tool as there is no support for external models.
Source: Google
