AI-generated faces that are hard to distinguish from the original

Recently, more and more active work is being done on artificial intelligence. And so AI produces all the best results, for example, human portraits are already difficult to distinguish from the pictures that artificial intelligence creates. But first you need to figure out how it all works?

How are faces created?

It is important to start with the fact that a deep neural network is used to generate a portrait. It is initially trained using a huge number of images. And thanks to a whole stream of data, you can get new opportunities in creating faces. In the end, it turns out to make really realistic portraits that are not much different from the original. Such images are called CAN faces.

Technology features

British scientists conducted a study in which they proved that faces created by AI have a negative impact on user behavior on the World Wide Web. First of all, this undermines trust. That is why some users are really distrustful even of real photos of people, not to mention CAN. But at the moment it is not established why everything happens this way.

Face for espionage

Not so long ago, there was a scandal in the United States over the fact that a CAN face was used as an avatar in one of the LinkedIn profiles. And the person who was hiding behind him could conduct dialogues, even with officials. And this case was not the only one. This kind of technology is used by attackers more and more often.

What to do with it?

Reality monitoring can affect this problem today. This is a method by which you can search for the source of information. For example, to understand whether it comes from the outside world or the fruit of someone’s imagination. In fact, AI is an opportunity for each person to train their own brain, and over time, they treat all information on the network more critically, not trusting fakes.