If you are on the Internet, it’s likely that you have already encountered a deepfake – whether you were aware of it or not.
What is a deepfake? The two combined words in the name help us break down the meaning. The word “deep” is a reference to the AI-driven deep learning models that are fed videos of someone’s face from a multitude of angles until the computer creates a virtual model of that face. “Fake,” of course, refers to the fact that these videos are not a real person. Think of it as a very convincing photoshop done with video.
People have used deepfakes for entirely innocuous purposes, such as inserting actors into movies they’ve never been in – seeing Nicolas Cage play Lois Lane in Man of Steel is pretty absurd albeit humorous to some. It’s all a deepfake.
Nonetheless, there is an inherent ethical issue in using someone’s image without their permission. But it’s all in good fun, right? No one is really hurt.
Not necessarily. Deepfakes are becoming increasingly easy for the average person to create. In 2018, desktop software called FakeApp was introduced to the market with all the tools needed to start creating deepfakes – provided the creator has enough images and videos of the subject individual for the deep learning algorithms to work.
Some of the uses of deepfakes are extremely problematic on both legal and ethical levels. For example, deepfakes are used to impose celebrities’ images on to pornography, extremely harmful to someone’s career and reputation. What if someone decided to do that to you or someone you care about? It wouldn’t be so funny then, would it?
An equally harmful use is perpetrated by politicians and their supporters making it seem as though someone has said or done something they never have, usually in the form of depicting their opponent as stupid or despicable. When combined with emerging technology that imitates and manipulates voices, it is becoming increasingly more difficult to trust the things you see – and hear – online. Obviously, this could have serious implications in the upcoming elections. If it is so easy to falsify video and audio of public figures, then how will voters know the truth? Indeed, one wonders what effects deepfakes have had on past elections. The technology has been around for quite some time and over numerous election cycles. Was that really your candidate?
It is said that a picture tells a thousand words. If that is true, then a video tells even more. So in a world where seeing – and hearing – is no longer believing, it is more important than ever to be aware of the content you consume and its sources or you may be the victim of a thousand lies.