The technology used to make anyone do or say things they've never done may seem complicated – but it's very achievable. All it takes is a few visuals or a recording of a voice to start creating an alternate reality.
This type of manipulated content that is created with the help of AI technology is called synthetic content, or more widely known as a deepfake.
Deepfake is a term that describes audio and video files that are created using artificial intelligence, "machine learning" to be more precise.
All kinds of “deepfakes” are possible. Face swapping, where one person’s face is replaced by another. Lip syncing, where a person’s mouth speaking can be adapted to a piece of audio that is different from the original. Voice cloning, where a voice is “copied” to use that voice to say things.
Completely synthetic faces and bodies can be created, for example, as digital avatars. With deepfake technology, even dead people can be brought back to life.
These synthetic video manipulations are produced with so-called Generative Adversarial Networks, or GANs.
A GAN is a machine learning model in which two neural networks compete with each other to become more accurate in their results.
In simple terms: A computer is telling another computer whether the digital clone it has created of you is convincing enough by comparing it to the original footage. Do you move the same, do you sound the same, is your expression the same? The system improves over time until it is satisfied with the result.
Although this technology is constantly improving and is very sophisticated, you can still spot deepfakes if you know where to look.
The full article published by Deutche Welle