Democracies will have to accept an uncomfortable truth: in order to survive the threat of deepfakes, they are going to have to learn how to live with lies.
Deepfakes are fake clips of people that look indistinguishable from real clips. The technology allows changing the face of a person, the voice, the facial expression, and even the body movements. Generative Adversarial Networks (GANs), a type of artificial neural networks, are used to create these clips. The technology which once was only available to Hollywood now is available for pretty much everyone with a laptop and some knowledge in the machine learning field. And soon it might be even more accessible when applications are developed to apply it.
Some researchers are working on tools to fight the disinformation and detect deep fakes. One method would be to detect unusual patterns in the picture, blinking or face movements. However, fake detection methods must evolve in parallel with faking technologies.
Check out this Bloomberg compilation of deep fakes and let your imagination run wild imagining all the potential evil applications of this technology:
This article is part of our development team tradition of sharing a fact of the day during standup.
Resource: Deepfakes and new disinformation war