Google takes on ChatGPT: its AI can now analyze videos and tell you how far back they were recorded.

ChatGPT brings together more than 120 million users a day, setting the trend for AI chatbots thanks to its built-in features that other artificial intelligences don't yet have. One of these is its advanced voice mode , which allows you to speak very naturally with it, or, more recently, its ability to create realistic, Ghibli-style images, something that has become a viral phenomenon.

One of its star features is its video analysis , which involves uploading a video and having OpenAI 's chatbot analyze it so it can be used to chat with you about it. This feature seemed to be only available in ChatGPT, but now it's been discovered that Google's Gemini AI can do it too .

Gemini can analyze videos like ChatGPT does.

Android Authority reports that they've discovered code within Gemini that allows video analysis, although it's only available in the beta version after configuring a few settings. They explain that, just like ChatGPT, to attach a video, you have to tap the "+" icon and select it from the gallery .

They add that the results are top-notch and that Google's AI is able to determine where the video was recorded by comparing it to its Google Maps database . As with the newly discovered feature of OpenAI's chatbot , which allows you to find the exact location of images, Gemini can also analyze the scene to give you a description .

Gemini can analyze videos like ChatGPT does.

The code itself reflects that multiple videos can be uploaded at once , allowing for separate analysis and even comparisons. As for video length, Gemini could have a five-minute limit , which would be a decent amount of time to use this feature.

For now, it's impossible to use without access to the beta version , and it's unknown when it will be available for all users. If it is, it remains to be seen whether it will be available in the free version or in the version exclusively for Advanced users . If Google is serious about competing with OpenAI, they may want to do the former.