We’re sure you have seen a video of someone that was manipulated. These seemingly innocent videos that started out as just pranks or for entertainment have been used more and more in mainstream media. The fact of which can be quite terrifying. Deep Fakes are AI-generated videos or synthetic media, sometimes even being more convincing than CGI or the original video.  

Artificial intelligence learns what a source face looks like from different angles and subsequently transposes the face onto someone else’s face as if it were a mask. This is not to be confused with filters. A filter is a technique through which size, colors, shading, and other characteristics of an image are altered. What makes Deep Fakes so terrifying is that it enables real footage of people to be manipulated.

According to Howtogeek: “It’s first trained with a target face. Distorted images of the target are run through the algorithm and it learns how to correct them to resemble the unaltered target face. When the algorithm is then fed images of a different person, it assumes they’re distorted images of the target, and attempts to correct them” 

One of the most infamous examples of a Deep Fake is a video of House Speaker Nancy Pelosi that was altered to make her look and sound drunk. It was widely distributed and thought to be real. Trump reposted it and has been viewed more than 3.5 million times on Twitter. Now, Deep Fakes have even been cited as a potential nuclear threat.

Perhaps the most vulnerable group of all, really are children. Experts have become better at spotting them, however, it is still incredibly difficult to spot even the most modern of Deep Fakes. Back in the day, it was a lot easier to spot a video that was altered. Technology and deep learning was not so advanced and you could easily spot inconsistencies with colors, contrast and movement. However, since a Deepfake literally maps the face of the target it has become increasingly difficult to spot. 

Check out this clip of Obama:

How do you spot one? 

As stated earlier, it has become increasingly difficult to tell the fake from the original. Experts however have noted a few things to look out for when spotting a Deep Fake. 

Initially, experts weighed in that Deep Fakes do not blink properly. Not too long after, they started blinking like normal people. It’s as though the moment any flaw is called out, they are corrected almost immediately. 

“Poor-quality Deep Fakes are easier to spot. The lip-synching might be bad, or the skin tone patchy. There can be flickering around the edges of transposed faces. And fine details, such as hair, are particularly hard for Deep Fakes to render well, especially where strands are visible on the fringe. Badly rendered jewelry and teeth can also be a giveaway, as can strange lighting effects, such as inconsistent illumination and reflections on the iris.” – The Guardian

So are they all bad?

This artificial technology can advance art, education and accessibility. One such example is Deep Fakes can bring historical figures back to life for a more engaging and interactive classroom. A synthetic video of reenactments or voice and video of a historical figure may have more impact, engagement, and will be a better learning tool.

The applications for this technology are endless. The technology isn’t going away. We can only hope that as a society we can use the applications for good rather than harm.

Interested in making one?

Your best bet is most likely with Adobe After Effects. Luckily, Innovative Technologies sells the entire Adobe suite. Why not try your hand at creating one…for good of course!

#Adobe #AfterEffects #AdobeProducts #AdobeInZimbabwe #DeepFakes