Although we have not yet come to robots in human form who speak like real people, artificial intelligence is making huge strides and truly capable of situations that are hard to believe can be fake and not real.
We have seen, for example, the voice assistants able to answer elaborate and self-educated questions or even how it is possible to manipulate videos of changing the faces of the protagonists and have a program that draws by itself.
Now a really interesting and special site has come out that using an innovative program automatically generates faces of people who don't exist .
These faces are not drawn as in a video game, but they are photographs that seem real, of imaginary people that nobody could define as false.
The thispersondoesnotexist.com website uses the algorithm developed by the NVidia StyleGAN project to show the fake photograph of a person who doesn't exist, with a face that looks incredibly real, every time it is opened.
You can update the site several times to always see new faces, all realistic, but false.
This is possible thanks to the StyleGAN algorithm which is trained by processing a huge dataset of real images and which then uses a type of neural network known as Generative Adversarial Networks (GAN) to generate new faces.
GAN is therefore a machine learning concept with the task of generating false images that are not distinguishable from the real ones.
You can train GAN to remember human faces, or even animals, cars, buildings and anything else you want.
The faces generated by the site thispersondoesnotexist.com are therefore fake photos that look real and that, theoretically, can be used to create fake profiles on Facebook or to actually create false identities that are almost impossible to distinguish from real identities.
A different site allows you to find thousands of faces of people who don't exist is //generated.photos/ . The photos are entirely generated by artificial intelligence and can be freely downloaded and used for any creative project.
We have seen, for example, the voice assistants able to answer elaborate and self-educated questions or even how it is possible to manipulate videos of changing the faces of the protagonists and have a program that draws by itself.
Now a really interesting and special site has come out that using an innovative program automatically generates faces of people who don't exist .
These faces are not drawn as in a video game, but they are photographs that seem real, of imaginary people that nobody could define as false.
The thispersondoesnotexist.com website uses the algorithm developed by the NVidia StyleGAN project to show the fake photograph of a person who doesn't exist, with a face that looks incredibly real, every time it is opened.
You can update the site several times to always see new faces, all realistic, but false.
This is possible thanks to the StyleGAN algorithm which is trained by processing a huge dataset of real images and which then uses a type of neural network known as Generative Adversarial Networks (GAN) to generate new faces.
GAN is therefore a machine learning concept with the task of generating false images that are not distinguishable from the real ones.
You can train GAN to remember human faces, or even animals, cars, buildings and anything else you want.
The faces generated by the site thispersondoesnotexist.com are therefore fake photos that look real and that, theoretically, can be used to create fake profiles on Facebook or to actually create false identities that are almost impossible to distinguish from real identities.
A different site allows you to find thousands of faces of people who don't exist is //generated.photos/ . The photos are entirely generated by artificial intelligence and can be freely downloaded and used for any creative project.
