Deepfake: From Just for Fun to Dangerous Technology with Dozens of Victims on the Internet

Deep learning is a type of machine learning based on artificial neural networks inspired by the human brain. The method is used to teach machines how to learn from large amounts of data through multi-layered structures of algorithms.

Deepfakes typically use a computer learning network called a variational autoencoder, a type of artificial neural network that is commonly used for face recognition.

 

Deepfake technology resembles the 21st century version of Photoshop,but is far more dangerous. It carries many negative implications as it continues to advance the field of artificial intelligence. Deepfake is often controlled by a group of cybercriminals who have the ability to convincingly mimic other humans. Additionally, this platform is used to create numerous manipulated videos that look very real.

In a recent incident, a group of cyber criminals created a fake video with the face of Rashmika Mandan, a South Indian actress, and the body of Zara Patel, a British-Indian girl. This video caused a wave of reactions in the Bollywood community.

When was it first discovered, and what did it do?

Deepfake technology first surfaced in 2017 on the online platform Reddit. Initially, it gained notoriety for swapping the faces of celebrities such as Gal Gadot, Scarlett Johansson, Taylor Swift, and others, primarily for creating explicit content. This face-swapping was accomplished through the use of a machine employing deep learning algorithms. One key element in this process was an AI algorithm known as an “encoder.”

What is an encoder?

An “encoder” serves as a crucial component in the deepfake process. It enables machines to discern commonalities between two faces, reducing them to their shared features by compressing the facial images.

With the aid of an encoder, erroneous images are superimposed to manipulate the source content, while another algorithm, called the “decoder,” reconstructs the face with the desired orientation and expressions.

What does Deepfake technology do to original videos and photos?

Deepfake technology adeptly alters and manipulates the voices and faces of characters within various videos, frequently with explicit content.

Deepfake usage?

Primarily, deepfake technology finds application in unsavory edits, notably within pornographic video content. However, there have been instances, especially during elections, where politicians’ video clips were digitally altered and disseminated with falsely attributed statements.

Prominently, even well-known individuals with significant public profiles have fallen victim to deepfake videos. An example includes former US President Barack Obama, who was featured in a deepfake video where he appeared to refer to Donald Trump as a derogatory term. This video garnered widespread attention.

Furthermore, even Mark Zuckerberg, Meta’s chief, was depicted in various deepfake videos making claims about having “total control of billions of people’s stolen data.”

An AI firm named Deeptrace was also identified, which had cataloged 15,000 deepfake videos by September 2019.

What do experts have to say about deepfake creators?

Experts, including Danielle Citron, a professor of law at Boston University, have voiced serious concerns about the detrimental uses of deepfake technology. Citron highlighted the fact that deepfake technology has been weaponized against women, and recent incidents like the one involving actress Rashmika raise further safety concerns.

What else can deepfake do besides creating videos?

Deepfake technology extends beyond crafting AI-driven videos. It can also generate entirely fabricated images from scratch. Additionally, it has the capacity to create fictitious profiles and characters, complete with fictional names and accompanying videos.

Who is creating deepfake videos presently?

According to a report from The Guardian, a wide range of individuals, from industrial and academic researchers to amateur enthusiasts, can create deepfake videos. Visual effects studios and even porn producers also harness this technology for the manipulation and creation of content.