Deepfake is a type of synthetic media in which a person’s likeness is superimposed onto another’s body or face to create an artificial, yet believable video or image. This technology is used widely for entertainment purposes, but has also been used for malicious purposes, such as the spread of disinformation.
Deepfakes (sometimes spelled deep-fakes) are created using a combination of artificial intelligence techniques such as machine learning and facial mapping. The creator of a deepfake typically begins by training a deep learning algorithm on a collection of images or videos of the individual they want to replicate. Then, the algorithm typically creates an artificial neural network to map the features of the subject’s face in the source images or videos to the target images or videos, allowing the creator to superimpose the subject’s face onto a desired body.
Deepfakes are becoming increasingly popular and accessible. While they have traditionally been used widely for entertainment purposes, such as creating celebrity look-alikes, deepfakes have been used to spread false information and for malicious purposes. In addition, the technology is being used to create other forms of synthetic media, such as audio deepfakes, which are used to create artificial and convincing audio recordings of people saying things they never actually said.
As deepfakes become more accessible to the public, it is important to remain aware of their potential uses. It is also important to remember that creating a deepfake without the express permission of the person being replicating is illegal in some countries.
References:
Fountain, Henry. “Can You Spot a Deepfake?”. The New York Times, February 26, 2020. https://www.nytimes.com/2020/02/26/technology/deepfakes-detection-spotting.html.
Deng, Jingwei. “How to Create a Deepfake”. TechJunkie, October 10, 2019. https://www.techjunkie.com/create-deepfake/.
