Last month AI’Twins“debuted with the ability to check if the image was created using artificial intelligence tools. Now they have gone even further and also can detect videos created with it.
Gemini’s AI video detection feature is very easy to use. The process is very similar to using the image detection tool. You just need to open the Gemini app on your phone or visit their website. then, We have posted a video that we believe was created by artificial intelligence and we asked him: “Was this video created with artificial intelligence?“.
However, the problem is that Gemini can only detect videos created with Google tools at the moment. This is the same limitation that exists when using their AI image detection tool. Also, right now we can only upload video with a maximum size of 100 MB and a maximum duration of 90 seconds.
However, it takes a few seconds for the chatbot to analyze both the images and audio of the uploaded video and tell us if one or both were created by AI. We’ll also see mention of the SynthID watermark in the results, which essentially helps Google identify AI content created with its own tools.
Twins
Twins
Generative AI can be used to generate images, videos and text. However, to make it easier to determine whether content is created by AI or humans, Google uses SynthID technology which adds a watermark to all content created with its AI tools. This digital watermark is invisible to humans but can be easily detected by a detector.
Essentially, when you ask Gemini if an uploaded image or video is created by artificial intelligence, it uses the SynthID detector to look for watermarks. If they are found, the chatbot reports that artificial intelligence has been detected in the uploaded content. Otherwise, you will receive a negative response indicating that the content does not contain any traces of artificial intelligence.
This watermark is only integrated into content created or modified by Google AI tools. So when uploading generated content with other AI tools like ChatGPT, Gemini cannot determine its veracity.
Companies are working to ensure that detectors can identify content generated by any AI tool create a Coalition on Content Origin and Authenticity (C2PA) as universal standard. As part of this effort, most AI-generated content will contain C2PA metadata.
However, it is still under development and many AI tools have not yet adopted the standard. Google recently announced that images created with Nano Banana Pro will be integrated with C2PA metadata.
With the development of AI tools, it is becoming increasingly difficult to distinguish real content from AI-generated content. We may think that this will not happen to us, but I’m sure we’ve already fallen into the trap. one more time.
We are not talking about videos made clearly for laughs, but about more dangerous ones, for example, of a political nature. Things They are not so easy to distinguish, especially for older peopleand very well-made videos are already starting to spread on social networks.
2025 gave us the birth of AI-generated images and videos. Let’s see what awaits us in the future.
