For many years now Google has defined herself as an artificial intelligence companynot as a search engine (which now has AI), an online advertising company or anything else for which it knows it mostly in countries like Spain. We must recognize that they had not only the vision to imagine that this was going to be a revolution, but to lead it by creating technologies such as the Transformers, that would give rise to all this tectonic change.
And this change is about to make a jump because The generation of videos with artificial intelligence is already available to anyone Through I see 2, a Google system integrated in Gemini Advanced, which is a version of payment of its artificial intelligence, but that does not imply an insurmountable wall since its price, from 21.99 euros per month, makes anyone use it.
We have been able to use it for a few weeks to check What are the limits of usemore in the quality and censorship that may have, than in the limitation of the number of videos, which occurs because the cost that Google has to assume due to the use of servers and resources is very high when it comes to creating videos and non -images or text.
The speed of change
The irruption of the chatbots has meant a change in the way we workwe entertain ourselves and even in which we try to take care of our mental health. It seems a lie that this is a reality in 2025, but it is. And it’s just the beginning.
Since the vast majority of large companies do not question that AI, in its multiple variants, will be the great change of this decade, at least, the economic and resources commitment is being huge. That makes advances are huge within a very tight time of timesomething we are not used to.
Google Gemini interface with I see 2
The free android
And this has a hidden face, and it is that People may not realize that it is totally plausible to create false videos that can go on authentic without the need for great infrastructure or, what is worse, advanced knowledge. Anyone can create a video from your mobile and, if you want, try to go through something real.
How a video is created in Gemini
The ease of use is one of the keys to this new Google proposal. Beyond paying for access. Gemini Advanced, You do not need to install any application or know how to program or design. You simply have to open the Gemini app on our mobile (or access the web version) and select the appropriate language model. At the top of the interface we must select the option I see 2which is the model designed for video creation.
New York video in view of Drone (made with AI)
Once done, we just have to give it a text -shaped command, trying to make it as detailed as possible so that the result seems to what we are looking for. The result obtained is always An 8 second video clip with 720p resolutionsomething that can limit the use that is given, although it is enough to illustrate other more extensive videos with explanations, tutorials or fictional stories.
We can create all kinds of videos, from some photorealists, such as the ones we usually find on stock video websites to illustrate certain situations, others more artistic or animation. We can also Create videos imitating certain styles. For example, with the command “A very sympathetic small unicorn that frozen for a pink meadow after a blue butterfly that flutters without paying much attention. The sky is full of clouds. Animation style similar to Pixar.” This is the result:
Vide of a unicorn (made with AI)
Nevertheless, Video creation is not yet as refined as imagesand there are still inconsistencies, especially with more elaborate commands. We can see artifacts (elements that should not be in the image) or animations that have no logical sense.
That is just what happens when We try to complex an order. In this case, the command used has been “a man drinking coffee in a very elegant Parisian cafeteria. Man has gray hair and after sorber of his cup contemplative via the window. The camera makes a pano around the scene to finally end up focusing the street, in which people are seen walking” and the result has been this:
A man drinking coffee (made with AI)
It is clear that We are in the early stages of a technology that is improving very fastand that it seems that it will not stop its advances for technical issues, beyond having to wait a few months, or perhaps a few years, to obtain videos that are indistinguishable from reality. Another issue is the limitations that the companies themselves put.
The limitations of see 2
Legal limitations, that there are, They are described by Google itselfbut there are also other more generic and subjective limitations. For example, They are not recreated naked even if they are requested since the video generated always hides certain areas so that they are videos suitable for all ages. For example, if a video is requested with the command “A naked woman lying on the bed looking at the ceiling” the result is this:
A woman in a bed looking at the roof (made with AI)
However, if you are looking for something more explicit, the message given by the system is this: “I cannot generate that video. Try to describe another idea. You can also obtain advice to write messages and review our video policy.” This limitation also occurs with videos that can try to simulate physical attacks and violence. Google knows that your tool could be used for illegal purposes and try to condition it to limit those uses.
For now there are failures in some videos or limitations imposed by companies such as Google, but also uncertainty because other companies can improve the system, not establish limitations and even train models to deceive human beings as best as possible. We enter an era in which we must doubt everything.
