Deepfakes — A Sorcery That Will Put an End to Our Society

We can no longer believe what we see — it’s not humans, it’s the computers.

Back when I was 13, the only thing that occupied my mind in the dead of the night was having someone like Forest Gump in my life — I don’t think people as nice as him exist anymore. Who wouldn’t want somebody who was met and praised by John F. Kennedy? The first time I watched this movie, I was convinced that Mr. Gump was my kind of a guy and that one day, I would say out loud how we are like “Two peas in a pod.” In my dumb days, I believed that my childhood crush met JFK, but little did I know — video manipulation was the spell that had started in the 1950s.

Deepfakes — What is it and who started it?

Deepfake is a type of video fraud that makes people look as if they are doing and saying things that seem impossible, controversial and sometimes, funny. Just like Photoshop has helped people with editing pictures long before Deepfakes became a thing, this technology not only eases the way to make this manipulation accessible but also makes it increasingly difficult to question its authenticity.

Deepfakes started with Generative Adversarial Networks (GAN). Back in 2014, GAN established a direct conflict between two neural networks called generator and discriminator. To make it simple for understanding, these two algorithms were to train each other. The first was meant to cover the counterfeit as real as possible, while the second tried to detect the forged. If GAN could have this many applications, then wouldn’t it be fair to say that GAN was the foundation of Deepfake technology — used to create, manipulative, eye-catching images, audio, and video?

The term "deepfakes" comes from a Reddit user in late 2017. In December, Motherboard found a Reddit user named “Deepfake” who got famous for his hobby of swapping celebrity faces onto pornstar's bodies. The scary part wasn’t that it was possible, it was the fact that all those videos were convincing and had people confused. Celebrities including, Taylor Swift, Gal Gadot, and Maisie Williams were the first few who were experimented on the deepfake’s side of pornography.

How is it used?

Deepfake's core element is machine learning. To create a fake video of someone, the maker first trains the neural network of the person's real video material for several hours to deeply understand it from different angles, shadows, and lighting conditions. The educated network is then combined with computer graphics technology to place a copy of the person on top of another actor. The person working, also needs to manually adjust many training parameters to avoid obvious jumps and flaws in the image — might sound easy, but it isn’t.

Unlike most facial recognition systems that transform facial features into computer unique codes, Deepfake's technology blends the identity with detailing we don’t even think it’s questionable. This is a nightmare that not only messes with your beliefs but also manipulates the general opinion of influential leaders, politicians, celebrities, and almost everyone.

It is spreading faster than light

Recently, some fake videos got their rage of virality and millions of people around the world have used this new technology. Who remembers President Obama mouthing profanities to describe President Trump? Bill Hader suddenly becoming Al Pacino in the late-night talk show, and a Christmas video where the Queen of England jokes about Harry and Meghan? These are only a few of them.

The number of Deepfake content on the Internet is growing rapidly. According to a report from startup Deeptrace, there were 7,964 Deepfake videos online at the beginning of 2019. After just 9 months, this number increased to 14,678. There is no doubt that it has continued to grow ever since. Today's Deepfake technology is very impressive, but still doesn't quite fit the real lens (has the potential). If you look closely, you can easily tell if the video is fake. But technology is advancing at an alarming rate, and it won’t be long till we believe things that aren’t real — just because of how they look.

It has also been a grab-me-rope and brought out many business opportunities. From a broader perspective, Deepfake is a type of synthetic media produced and modified by AI and has created its place in a huge market. A few months ago, Samsung Next Ventures plotted a landscape map in a whitepaper. And they believe synthetic media is the beginning of the third phase of media development.

Deepfakes spelled backward is catastrophe — it has wrecked our society

The Deepfake application of swapping faces is free and easily available over the Internet. So for anyone who has time, simple technical knowledge, and a powerful computer — can easily access and get their hands dirty in this technology. Other types of deep fakes are more complex. Fortunately, this means your risk of falling as the victim is lower — But wait, they are filled with more dangerous things and tools.

Yes, this kind of manipulation is not new. However, advances in artificial intelligence, easier access to video editing tools, the speed and freedom of this technology are enough to put an end to our society. Not long ago, the controversial web application DeepNude showed how to use this AI technology and allowed users to create real nude women's pictures, simply by uploading face photos to the app. It roared high, and many people use it as a source of blackmailing and manipulation. The application was later shut down — but it definitely ruined many lives.

Bottom line:

There are many opportunities birthed because of this technology, but the danger to humanity is worth more than any opportunity. Before our enemies make people believe things against us, our economy, and our society — there should be an end put to this sorcery.

Seeing is believing — the times when Artificial intelligence and Machine Learning were considered safe for everyone — now, it is not the same. This manipulation needs to stop, and we need to look out before it messes with our personal lives.