Artificial Intelligence: Deep Fakes and Fake Faces
How would you stay if I told you that some people on the internet are created by artificial intelligence? They are false faces of people who do not exist! Without a doubt, it seems straight out of a science fiction movie, but surprisingly it is already part of our reality.
Since “Snapchat,” the app that went viral for its 24-hour format before Instagram launched “Instagram Stories.” At that time, Snapchat launched its face swap filter. Do you remember? It’s scary how everything evolves.
The evolution of fake faces: What are “deep fakes”?
Recently, a video of actress Jennifer Lawrence receiving an award and giving a speech went viral on networks. Well, what went viral was the actress giving the speech but with the face of actor Steve Buscemi. What madness! This video became a clear example of how fake news is no longer just text but fake faces, better known as Deepfakes.
Deepfake technology allows you to create completely fake photos from scratch. But, there have been cases like “Katie Jones,” a profile that claims to work at the US Center for Strategic and International Studies but was actually created to carry out espionage operations abroad.
How are deepfakes created?
The “deep fakes” are created through artificial intelligence, which works through an encoder that takes thousands of shots of the face of the people who want to fake it.
The function of this algorithm is to identify and learn the similarities between faces and reduce them to shared common characteristics through an image compression process. Then a second algorithm called a decoder comes into operation, responsible for recovering the compressed images.
To carry out face swapping, the encoded images are entered into the opposite decoder so that it reconstructs the other person’s face. That is, an encoded image of face A is fed to the decoder trained on face B.
The decoder is in charge of reconstructing face B with the expressions and orientation of face A to obtain a clear and super realistic image.
Fake But Realistic Faces: The Artificial Intelligence Revolution
But, obviously, human beings were not going to stop there. Artificial intelligence must evolve much more. Artificial intelligence can create an entirely false face portrait of someone who does not exist.
We are talking about the “This person does not exist “website; it is a new demonstration of how easy it can be to create “personas” that are not real. Its creator, Phillip Wang, is a software engineer fascinated with GANs. His main objective was to set up a site where he could show the capabilities of the GAN: fake faces.