In 2017, an anonymous Reddit user under the pseudonym "Deepfakes" posted several porn videos on the internet that appeared to contain a series of celebrities including Emma Watson, Katy Perry, Taylor Swift and Scarlett Johansson. These were all eventually debunked as fake.
The origins for this phenomenon can be traced to 2014 and a graduate student, Ian Goodfellow, who invented a way to algorithmically generate new types of data out of existing data sets. Three years later, and the software system is capable of sourcing video content on the web and face matching it to pretty much any pre-shot footage. The combination of the existing and source videos results in a fake video that shows a person or persons performing an action at an event that never occurred in reality.
The creator of FakeApp, a face-swapping video app, says he wants to improve the app to a point where users can simply select a video on their computer, download a neural network correlated to a certain face, and swap the video with a different face with the press of a button. Scary stuff!
This technology has now spilled over into, even more worryingly, ‘fake news’. With the power to be able to skew information, manipulate beliefs, and push extreme ideologies using a whole library of politicians and celebrities both alive and dead now a reality how can we protect ourselves from this possibility?