What is deep fake: technology that can make you believe anything?

Artificial intelligence has gotten far enough to be extremely easy to trick you into. Deepfake technology takes everything to the next level.

What is deep fake: technology that can make you believe anything?
What is deep fake: technology that can make you believe anything?


A video with Mark Zuckerberg, Facebook CEO, appeared on Instagram. Here, he recognizes that he has control over all data, secrets, and even the lives and the future of billions of people. Users instantly realized that the video was a fake video and that it was actually a deepfake. Still, images seem real enough for some to fall into the net. For this reason, I think it's good to know a little more about this technology.


Deepfake technology has been in place for some time, and the name refers to both the method by which they are made and the resulting images in the process. It is an adjunct to two terms, deep learning (the technique used in artificial intelligence that helps computers learn through examples) and fake, which you can see for yourself.

Thus, artificial intelligence allows anyone to create a compelling simulation of a human subject. Everything is possible through deep learning, technology by which the computer is shown enough images with that person and then uses it to create these fake videos.

The worst part is that the pictures seem so real that you can easily fall into the net.

This video, as well as other deepfake with Donald Trump or Kim Kardashian, were made for an art exhibition. Its purpose was to draw attention to the way people can be lied, manipulated, or even monitored by politicians, celebrities, or influential people.

In the case of deepfake with Mark Zuckerberg, the advertising agency that created him announced that he used an original video with the Facebook CEO. He thus trained an AI algorithm with those initial images and an actor's audio recording, which is heard in the background. Therefore, he rebuilt the frames to make Zuckerberg's facial movements match the rhythm of the actor's voice. In fact, this was the reason why most people realized it was a fake: the sound did not match that of the Facebook head.

The advertising agency says it created these deepfake to educate the public, who should better understand how artificial intelligence can be used.

Why the video will not be deleted from Instagram
These videos come as a test of Facebook's policy regarding the fake images on its platforms.

The fact that Facebook's CEO himself was involved in this situation raised the following question among the people: Will the video be deleted from the social network? The answer seems to be that it does not. The company has announced that it will treat the video just as it would in any other form of misinformation on Instagram. Usually, misinformation is not deleted from social networks but is not recommended to users.

The tech giant might have reacted altogether if something similar would not have happened earlier this year. Facebook refused to delete a fake video with Nancy Pelosi, an American politician. It had been modified to make Pelosi drink. At that time, the company had to define its position regarding the altered images. Decided then your video will stay on the platform.

This complicates any attempt by the company to delete the video with the Facebook founder. He would be hypocritical in the process. If he had predicted the situation, he might have done otherwise with the American politician.

How easy it is to do that?

The exciting part is that these deepfakes are not very difficult to accomplish. Researchers have recently demonstrated how software can be used to make a subject say what you want.

Thus, you can edit a video clip to add, delete, or edit the words that come out of a person's mouth. The results were entirely satisfactory. From a group of 138 volunteers, approximately they considered deepfake videos to be real.

Although artificial intelligence is an incredible technology that makes life easier in many ways, it also comes with negative parts. With this deepfake, the fake news phenomenon could grow. It will be easier for those malicious to promote false content as truthful.

Soon you will not know what is right and what is false, which makes it even more necessary to use the internet with care and a skeptical approach to all the information running on the web.

A good explanation of the danger that deepfake represents is provided, well, even though a deepfake. Actor Jordan Peele has used artificial intelligence to deliver a kind of public service announcement about the dangers deepfake technology poses. He did this with the image of Barack Obama, the former president of the United States.

In the case of such videos where the subjects say or do something out of the ordinary, it is advisable not to take everything well first. Check the source and look closely at the images to see any sign that they are false (usually these modified videos contain deformations and blurring of the vision).

Perhaps the phrases pronounced by the CEO of Facebook are not far from the truth, but more positively, Mark Zuckerberg would never have said this publicly.

Experts are now working on tools where fake can be identified, but until they become available, the best protection is a better selection of the information we encounter in any form they would present to us.