deepfake

Fake News Vs Deepfake

Recently we have heard a lot of talk about fake news. Fake news are false news published to get more clicks and increase website traffic or to harm someone´s reputation.

But…What is deepfake?

Deepfake goes a step further. In this case, we are no longer talking about false news, but false videos or audios created with artificial intelligence (deep learning). Through the use of an automatic learning technique (GAN) and real data, a false version  of a video or audio is created. The video itself is original, but the face of the person IN it is not. The manipulation is so effective that is not perceived by users.

Anyone can be the victim of a deepfake video

A well-known case of this type of scam is the Barack Obama´s video. The University of Washington created the images from audios of the ex- president. In the video the former president came out saying that Donald Trump was a complete idiot. The key in these cases is to achieve a prefect synchronization between the audio and the movements of the person’s mouth. If we pay attention and observe carefully, in many cases it is possible to notice that the movements do not fit perfectly or that the person in the video does not blink or behaves in an unnatural way. However, in those cases in which only the eyes or mouth are changed, it is more difficult  not to fall into the trap.

What are the risks?

Given the amount of information to which we are exposed and the ease with which social networks (based on sharing, commenting) can make something false appear to be legitimate, we need to be very cautious. «On a personal level, it can cause us harm, but applied to polarizing areas such as politics or economics, the stability of an entire region could be compromised.  Or «…the stability of an entire region is at risk or could be at risk.

Is it possible to detect it?

Yes, the same technology used to create them is the one that can best help us detect them. Through artificial intelligence we can apply deep learning to detect deepfakes automatically. The more a counterfeiter evolves, the more his ability to detect counterfeits does. Beware of oversharing. To think that we can put any word in the mouth of anyone is quite scary. What would happen in a world in which the line between the real and the false would be blurred and we could not trust the information?

Follow us on Twitter y LinkedIn to stay up on the latest updates and remember that you can also subscribe to our blog!.