If you’re active on social media, you’ve probably heard about deepfakes. Deepfakes are videos, audio clips or photos that are created using artificial intelligence, but they appear completely real. Many people make these for fun or parody, but this technology is also being used for scams. Scammers are now creating deepfake content which people consider to be genuine, fall into their trap and suffer losses. On a larger scale, deepfake content is also being used to influence elections in many countries. In today’s explainer, we will learn what deepfake technology is, how it works, and whether it has only disadvantages or some advantages as well.
What is deepfake technology?
Deepfakes refer to media, including photos, audio and video, that have been created or altered using AI. The term deepfake is a combination of the deep learning concept and the word “fake”. Tools for editing images, video, and sound have been around for a long time, but deepfakes are a new development. It uses machine learning to alter media in such a way that it becomes very difficult to distinguish between real and fake.
when did it start?
The term deepfake first came into the public domain in 2017 when a Reddit user named “deepfakes” shared edited pornographic videos on the platform. It used Google’s open-source deep-learning technology to replace the faces of performers in pornographic videos with the faces of celebrities. This is how the world heard the word deepfake for the first time.
How does deepfake technology work?
Deepfake technology has completely changed with the advent of AI. Existing deepfake AI works with two machine learning models that work against each other. It involves a generator algorithm that uses sample audio, video or images to generate or edit new content. This algorithm tries to create content that is most similar to the sample. Then, the discriminator algorithm comes into play. It is designed to identify deficiencies in the sample. This points out gaps in the content where the generator algorithm has made mistakes. To correct these deficiencies, the content is sent back to the generator algorithm. It is called Generative Adversarial Network (GAN). In this entire process, the generator and discriminator analyze data from the media sample. The generator then creates or modifies the media. This is an early deepfake. The discriminator then identifies the differences between the sample and the deepfake, and points out the shortcomings. The media is then sent back to the generator, which corrects those deficiencies. This process continues until the deepfake looks completely real.
second way
Another way to create deepfakes is to use an AI algorithm called an encoder. It is used to change or replace faces in videos or photos. In this method, a face is superimposed on a completely different body. Many deepfake tools also use two autoencoders, which allow both image and movement to be transferred from one image to another.
Disadvantages of deepfakes
Scams – Cyber criminals are using this technology to commit scams. For example, scammers are creating deepfake videos of people being kidnapped and demanding ransom from their families. Deepfake technology is also being used to carry out many other types of scams, due to which people are facing huge losses. Porn Videos: Many people are using deepfake technology to superimpose the faces of celebrities and common people on porn videos. This is a big threat from deepfakes and it can completely ruin anyone’s reputation.
Election manipulation: Deepfake technology is also being used to influence elections. During elections, a deepfake video can be created showing a politician saying something they never actually said. This danger is even greater in democratic countries like India. A few years ago, a deepfake video of Barack Obama went viral across the world.
Social engineering attacks: Deepfake technology is also being used in social engineering scams. In these scams, people are duped using deepfake audio. Scammers call employees of a company, posing as the CEO or other executives, and trick them into transferring large amounts of money to their accounts. Many such cases have come to light.
Spreading misinformation: Deepfake videos are also being used to spread false and misleading information. Some time ago, a video of Mark Zuckerberg went viral in which he claimed that he has complete control over the data of billions of people with the help of an imaginary organization shown in a movie. Many such conspiracy theory videos are going viral on social media.
Are there any benefits of deepfakes?
Deepfake technology can also be used for good. It can be used to create parody and satire videos that can entertain people. Additionally, deepfakes can be used for technology demos, recreating historical events, and so on.