This competition occurs during the data training process, where one algorithm generates data (the generator) and the other simultaneously discriminates or classifies the generated data (the discriminator). GANS distinguish themselves from other deep learning algorithms because they utilize two neural net architectures which compete against each other. Deepfakes rely on a particular form of deep learning called generative adversarial networks (GANS), which were introduced in 2014 by researchers at the University of Montreal. Deepfakes require a specific type of artificial intelligence (AI) and machine learning called deep learning, which utilizes artificial neural networks to become smarter and better at detecting and adapting to patterns in data. This was especially the case when President Donald Trump retweeted House Speaker Nancy Pelosi (D-CA) in a deepfake video, apparently slurring words, or when a troll posted a deepfake of Mark Zuckerberg giving a menacing speech.īefore analyzing the legal and social ramifications of deepfakes, it is important to understand how they work and why they are so effective. Equally appalling as the creation of these videos is their proliferation - and in their short history, deepfakes have made waves in the mainstream media. Other apps such as FakeApp generalize this experience to broader contexts such as bullying, blackmail, fake news, and political sabotage. Deepfakes thrived in this context, inciting the establishment of apps such as DeepNude, which offered both a free and paid service for altering images of women to appear completely naked. The first appearance of these convincingly-fake videos was in December of 2017, when an anonymous user on reddit under the pseudonym, “Deepfakes,” superimposed celebrity faces onto pornographic videos. This convenience poses new and unchartered legal and ethical challenges. Realistically altered videos and images once required high powered and expensive software combined with extensive training, whereas the creation of a deepfake is available to virtually anyone with a computer. While advanced digital alteration technologies have existed for years, they have largely been limited in access by cost and skill. What sets deepfakes apart, however, is their scope, scale, and sophistication. At first this might sound common and benign - Adobe’s Photoshop has been available for decades and movies like James Cameron’s Avatar wholly relied on computer generated imagery (CGI) technology.

The term “ deepfake” refers to an image, video, or audio clip that is doctored using deep learning to depict something that did not really happen. The year 2017 marked the arrival of the “deepfake,” and with it yet another new domain of technologi cally-induced legal ambiguity.

New deepfake app reddit