Deepfake has been a worldwide sensation for a few years now, so you must have certainly heard of it. Deepfake is a term that refers to the type of technology used to modify an image, voice, video, etc. to make it appear to be something or someone it is not. The term uses a combination of words: “deep”, as in deep learning or teaching artificial intelligence, and “fake” which means false.

The most popular use of deepfake is the alteration of a person’s identity. Basically, the face of one person is replaced with the face of another person using a technique called face reenactment.This means that facial characteristics such as skin color, hair color, gender, and/or age can be altered to create a new identity. There are even more types of deepfakes such as face merging, which combines two images to create a completely new one (look what our child would look like!), audio to video, which can make it look like someone is saying something they never actually said, or speech to speech (e.g. using different languages).

Unfortunately, when it comes to deepfakes, they can be used in many different ways. There are two sides to the story; one is a “good lie” and the other a “bad lie” as the presenter says.

Good lies are, for example, movies and perspectives where deepfakes are used to create amazing special effects and are practically everywhere. In “Lord of the Rings”, the trick of “forced perspective” was used to show a difference in height between humans and hobbits. Actors stood at different distances from the camera, and the one further away appeared smaller. We, humans, accept this lie. Why do we? Because it is entertaining. The presenter also gives several examples of this, so be sure to check them out.

Where can we also use the “good” deepfakes? Boom, gaming world. Indie games are able to thrive as only one person can create the voice of all the characters (voice deepfakes). Another example is culture. Wingardium Leviosa! Are you floating excitingly? 😊 Maybe in Harry Potter’s world they also used deepfakes to make the photos move… Besides bringing photos and works of art to life, technology can also help us improve, refresh and repair them.

So, what are the dangers of deepfakes? What is a “bad lie”?

One type of harm is a “social lie“, which is essentially the manipulation of news media. A deepfake video that passes for real news can be used to spread misinformation and can have real-world consequences.  Another one is “financial lie”, which is the use of deepfakes to facilitate identity theft, fraud, and extortion. Thieves can use technology to create convincing videos and images that can trick people into revealing sensitive information. And finally, there is the “psychological lie” that can lead to defamation, intimidation, and damage to confidence. We can spot bullying based on deepfake technology simply by logging into Twitter and searching for “Millie BobbyBrown”.

How to detect deepfakes?

Well.. one artificial intelligence against the other artificial intelligence. There are programs designed to detect lies.Researchers are aware of the risks deepfakes bring to the table, and the systems are able to detect manipulation of the images and technology. They look for evidence in the form of artifacts (distortions) because the hell of deepfakes is a moving face – and a fast-moving face is a death to it.

The webinar included an exercise on spotting a deepfake. Check it out and let us know how it went for you!

To sum it all up, what can we do? To protect ourselves from deepfakes we need to:

• Talk more about the penalization of deepfakes
• Be aware that deepfakes exist
• Be aware that we are susceptible to disinformation
• Engage in critical thinking
• Learn how to spot manipulation.

And finally, a beautifully crafted joke at last: why did the chicken cross the road?

To fact-check, if the other side had more reliable sources of information! Please verify your sources, even if you are just crossing the road.

Psst. We wrote about deep-faking a voice last month on our LinkedIn. Click!

‍