“I thought there’s no way this is going to work,” says Tal Hassner at Facebook. “How would we, just by looking at a photo, be able to tell how many layers a deep neural network had, or what loss function it was trained with?”
But it has indeed been made possible according to the claims that Facebook is making. The claim is that Facebook now has developed artificial intelligence that is very much capable of identifying the DeepFake images and that they are also capable of tracking down their Creators using the reverse algorithm technique.
DeepFakes are completely fake IDs that had been made from scratch. Facebook Artificial Intelligence is going to look for similarities within the identified DeepFakes too, just to make sure that the fakes are coming from a same-origin or not. It looks for those unique patterns that are very minutes like mole patterns or freckles. Such patterns are invisible to the human eye.
This is indeed a true step forward to technological development. “What we’re doing is looking at a photo and trying to estimate what is the design of the generative model that created it, even if we’ve never seen that model before,” says Hassner.
Artificial Intelligence has been tested too. The test was on 1,00,000 DeepFake profiles created. The end result was very good. “It’s a big step forward for fingerprinting,” says Nina Schick. She is the author of Deep Fakes and the Infocalypse. It is like a cat and mouse game between the developers and the people who create the DeepFake accounts. It is very interesting.