Theres a new trend on the interwebs called ‘Deepfakes’, a machine learning system that can be trained to paste one person’s face onto another person’s body, complete with facial expressions. The app uses the machine learning tool Tensor Flow from Google and has surprisingly good results. By doing so, a broader public became aware of the possibilities machine-learning algorithms offer for such face-swaps.
How much potential DeepFake has?
Using this system, users can edit videos, deceptively recreating real movie sets or share the faces of pornographic actors with others.Deepfakes may be used to create fake celebrity pornographic videos or revenge porn.
With “FakeApp”(More details below) tool, film and video scenes can be manipulated in such a way that a person’s head can be mounted on the body of a character or actor in a film, or a virtual film set can be recreated. For example, a Reddit user recreated the Leia scene from the new Star Wars movie. Instead of a $ 200 million budget, he uses the freeware FakeApp.
This was already possible earlier, but quite elaborate and optically quite easy to make out as “fake”. Meanwhile, however, an algorithm has been developed that with the help of artificial intelligence and machine learning – the algorithm gets better, the more often it is turned away deceptively real, for example, faces of other people projected on the actor.
Deepfake pornography surfaced on the Internet in 2017, particularly on Reddit and has been banned by sites including Reddit, Twitter, and Pornhub. Deepfakes can be used to create fake news and malicious hoaxes.
A desktop application is developed by a redditor called “FakeApp” which uses google TensorFlow library. The app simplifies the process of face swapping in videos which can be shared on internet.The app uses an artificial neural network and the power of GPU(graphics processing Unit aka graphic card) and three to four gigabytes of storage space to generate the fake video. This app uses the open source AI Framework TensorFlow developed by google, which among other things was already used for the program DeepDream.
Deepfakes wants to refine its software in the coming weeks. “Users should be able to choose a video on their computer, run a face-to-face neural network from a database, and then get a new video with their faces changed,” said anonymous developer Motherboard journalists .