Some time ago a few apps to put your face in movies, video clips, etc. became very popular.

Several videos of politicians and actors saying and doing things that, of course, they had neither said nor done went viral.

Today we talk about the technology behind these “virtual manipulations”: deepfakes.

What are deepfakes ?

Deepfakes are realistic representations of events that did not occur, and are artificially constructed.

People’s faces, their voices, their movements, their speech, etc. are manipulated, so the content of the video, photo or audio in which they appear changes.

How do deepfake s work?

Literally deepfake means deep learning + forgery. Therefore, it is about virtual representations based on artificial intelligence and machine learning.

This combination allows to obtain almost identical and very realistic results, produced entirely by computers (CGI).

It seems simple, but to build a fake video of a person, huge amounts of photos, videos or audios of the person in question have to be collected. From all this information, it is when you can create standard models and formats capable of recognizing faces, movements, gestures, sounds, etc.

Similar to the technology behind the facial recognition, once trained, these systems are capable of replicating in the form of deepfake all the data they have stored.

Of course, if a person is not exposed to the media, it is very difficult to get the amounts of information that are needed to build a deepfake . Hence, those most affected by this problem are people with a public life, such as actors, singers and politicians.

Deepfakes and identity theft

Identity theft is a serious problem that has been increasing in recent years.

Deepfakes are directly related to this fraud, since they are a form of identity theft.

The majority of people – celebrities and politicians- that appear in this type of format do so without your knowledge. They are victims of identity theft because they totally lose control over their image, which can be affected.

Can an identity verification system be fooled with deepfakes?

Short answer: No.

While there is fear that a fraudster may be able to build a video falsifying the identity of a person to pass identity verification processes. The truth is remote identity verification solutions are based on anti-spoofing techniques, capable of detecting deepfakes or any other type of phishing fraud. In fact, this is done by the liveness detection. And a user performing this process will not be able to authenticate if they have not previously passed this filter.

In addition, identity verification solutions based on biometric systems, as in our case, they involve very sophisticated security algorithms that are capable of recognizing if a person is real and is taking a selfie in real time or if it is a manipulated video.

Our digital onboarding process is carried out in real time through a video-selfie capture and identity documents at the time of verification. At all times, the technology is able to recognize if the video has been recorded in advance.

As if that were not enough, our technology not only focuses on the individual, but also takes into account background, lighting, depth, etc., ensuring a thorough check-up.

And we must remember, as we mentioned above, that people not exposed to the media do not run great risks because it is very difficult for fraudsters to have access to large amounts of their audiovisual information.

Good uses of deepfakes

The popularization of deepfakes has raised a lot of criticism and concerns due to all the fake news and malicious information that has emerged.

Like all technologies, they give rise to great ethical debates and doubts about their applications. However, we must not forget that it is a tool and, therefore, its result depends on who uses it.

So deepfakes offer a variety of options, for both profitable and fraudulent purposes.

In this section, we delve into the most prominent examples of good uses of this tool (which, of course, has a lot of potential):

History through their main characters

The educational field could take advantage of the full potential of this technology to revive historical figures and tell their own trajectory, answer questions, interact with classes, etc.

Definitely, deepfakes can offer interactive lessons. Besides, this opportunity could also be transferred to museums, libraries, etc.

Better training and practice

In the professional world, deepfakes can also be used for realistic practice. For example, trial job interviews, rehearsals of customer interactions during trainings, oral exam preparation, etc.

Realistic dubbing

Deepfakes can be used to translate films, documentaries, etc.

With this technology, it is possible to replicate the voice of the actors without having to go through the process of dubbing an entire film in another language. Of course, adjust the translations to the gestures and movements of the lips, obtaining much more realistic results.

Aging and rejuvenation of characters

Video manipulation has multiple uses in the entertainment industry. Specifically, in films that reflect various stages in a person’s life, deepfakes allow to maintain consistency in the appearance of a character (and avoid problems in casting selection because similar people are not found).

This technology can be also used to replace actors who, for various reasons, cannot continue a recording.

People’s privacy in sensitive environments

To avoid putting the privacy of real patients at risk, deepfakes can be used to maintain the privacy of people in studies and research.

Now that you have more context, what do you think of deepfakes? We read you at LinkedIn !

If you liked it, share it on