Seeing Through The Illusions Of A Fabricated World

January 8, 2021

"On the internet, nobody knows you're 
a dog"

A dog sitting at a computer in Peter Steiner’s 1993 New Yorker cartoon apprehended a revolutionary change of human interactions. We've all learned the dog's lessons over the past century.

A random person on the internet could be anybody. The fifteen-year-old kid in a chatroom could be an undercover cop or a spy who wants to infiltrate the intelligence community. The gorgeous lady on Tinder could be a wife trying to catch her womanizer husband. The business consultant from a famous establishment who offered you a business proposal could be a professional scammer.

We may know how to identify a scammer's name and messages, but pictures are different. A scammer could use someone else picture, but easy to identify with the use of Google reverse search. So sometimes, we tend to trust the picture. And there you are, being fooled by your sexy and gorgeous match on a dating site who is actually 10 pounds heavier than what is on the picture. The American Navy who portrayed the same age as yours is actually 10 years older.

In this new era with lots of innovations, fantasies can be turned into a reality. If you are looking for characters for your websites to make them more diverse, there are a lot of websites making generated photos that you can use for free. New machines learning algorithms allow people to quickly generate artificial 'photos' of people who have never existed.

They just need to adjust the likeness they needed; make them old or young, sexy or fat, or the ethnicity of their choice - American, Australian, African, Asian, European, etc. A company called Rosebud.AI can also make your fake person animated, even make them talk.

Do they look familiar to you?

Perhaps you saw these faces on Instagram, Twitter, Tinder, Facebook, Tiktok, and online posts and advertisements. These generated people are starting to show up on the internet. The creation of facial fakery becomes possible with the help of the new type of artificial intelligence called a Generative Adversarial Network (GAN).

To do this, a lot of photos of real people are feed to a computer program. There are two main parts of the system: the first part studies the photos of different people and tries to come up with its own photos, while the other part uses to detect fake photos. The end product is indistinguishable from the real photo.

The A.I. system recognizes each face as a complex mathematical figure, a range of values that can be shifted like those that determine the size and shape of eyes — can alter the whole image.

Source: www.nytimes.com

This idea was created by The Times using GAN software and released to the public by Nvidia, a computer graphics company. This system gives us difficulty to identify who is real online and who is only a creation of artificial intelligence.

The technology becomes so advanced at identifying facial features. Our smartphones can now use facial recognition to unlock them, and even search a certain person through the thousands of photos. Law enforcement is now using facial recognition to identify and arrest criminals. Even some activists are using facial recognition to reveal the identities of police officers who are trying to remain anonymous just to infiltrate their groups.

Another company called Clearview.ai created an app to recognize a stranger using one photo. It scraped billions of public photos shared online to help law enforcement track down perpetrators and victims of the crime, including terrorists, human traffickers, pedophiles, and the most dangerous criminals. It aims to process the world in a way that was not possible before.

However, the same with other artificial intelligence, facial recognition algorithm is not perfect.

For instance,

• In 2015, Google's image-detection system labeled two black people as 'gorillas'. Maybe because the system has been feed with more photos of gorillas than dark-skinned people.

• Robert Williams, a black man in Detroit, was arrested for a crime he never committed due to an inaccurate facial recognition match.

Artificial Intelligence makes our lives easier, however, it is not perfect as we are, because it is only human who created it. We design a computer program, we feed it with data made by us, and we choose how the A.I. system will work.

Computers are truly good, but our visions are even better. If your eyes are good, you can spot fake things and people at a single glance. There are many websites like whichfaceisreal.com that were created to give awareness to people of which digital identities can be faked and help spot the fake parts at a single glance.

For example,

• Earings often not exactly match.

• Eyes may be the same distance from the center if it has been centered, cropped, and scaled.

• Glasses' end pieces may not match.

• Deep indentations in one ear that is not present in the other.

• Sometimes, artifacts appear out of nowhere that is hard to spot at first glance.

• Abstract or blurry background can be a clue to spot a fake person in the photo.

All photos from: www.nytimes.com

Computer software and hardware used to generate fake photos will continue to improve. So there will come a time that humans will fall behind the sides between forgery and detection. And we humans place trust in these kinds of systems, however, they can be as flawed as we are.

For us not to be a victim of these fake people online, we must learn how to spot and filter the fakes from the real ones.

Thanks for reading @Jane

Lead image from icons8.com