According to Wikipedia, deepfakes are synthetic media in which a person in an existing image or video is replaced with someone else’s likeness. The phrase is a portmanteau of “deep learning” and “fake”.
While the act of faking content is not new, deepfakes leverage powerful techniques from machine learning and artificial intelligence to manipulate or generate visual and audio content with a high potential to deceive. The main machine learning methods used to create deepfakes are based on deep learning and involve training generative neural network architectures, such as autoencoder generative adversarial networks (GANs).
Deepfakes have garnered widespread attention for their uses in celebrity pornographic videos, revenge porn, fake news, hoaxes, and financial fraud. This has elicited responses from both industry and government to detect and limit their use.
Deepfakes are important because they push the very edge of our ability to determine the credibility or accuracy of a piece of audio or visual content. While it’s easier to determine the veracity of text, altered photos, audio, and videos prove much more difficult to spot.
At Which Face is Real, we’re asked to decide which photo is real. It’s harder than you think. See below.