Deepfake: A Double-Edged Sword In Cybersecurity

A new term "deepfake" has become very familiar nowadays. With the assistance of Artificial Intelligence (AI), fake videos and images are being created and spread through social media. Many celebrities have already fallen victim to this cyber trap, including Bollywood actresses like Rashmika Mandanna, Alia Bhatt, Priyanka Chopra, and even music star Taylor Swift. Consequently, a sense of fear has emerged among everyone regarding the misuse of this technology. This fear is due to the manipulation of someone's body movements and voice in deepfake videos, making it difficult for many to distinguish between real and fake content. This confusion has led to instances of harassment.

Recently, Forbes published an article about deepfake phishing. Phishing has long been a familiar method in the cybercrime world. According to Forbes, phishing remains one of the most effective methods for hacking or intrusion. Hackers adapt their skills with technological advancements, and in recent years, phishing has taken on a more extensive form. As a result, some experts have labeled AI technology as the most dangerous form of cybercrime.

What is Deepfake Phishing?

Deepfake refers to synthetic images, videos, or audio created using deep learning. Hence the name "deepfake." This technology can be used through AI to alter photos or videos using Photoshop or mobile apps. It can change faces to make someone appear older or younger, clone voices, use scenes from movies, or manipulate facial expressions of a celebrity in inappropriate videos. The person who watches the video may mistakenly believe it to be real. Deepfake phishing is a relatively new form of phishing, where cybercriminals use social media and deepfake technology to attack or target individuals.

How Does it Work?

Similar to how someone is targeted on social media, deepfake phishing follows the same principle. Such videos confuse or manipulate people, exploiting their trust to obtain information. Cybercriminals use various methods, not just videos, to utilize deepfake as a weapon. For instance:

Email or Messages

Many organizations suffer from Business Email Compromise (BEC) attacks or CEO fraud, losing millions of dollars every year. Deepfake makes these attacks even more dangerous. Hackers use this method to make emails or identities seem trustworthy. For example, they create fake LinkedIn profiles of CEOs. Then they use these fake profiles to send messages or emails to employees.

Video Calls

Cyber attackers can use deepfake videos during Zoom calls to steal sensitive information or even consent to transactions. Because the person on the other end of the call cannot detect the deepfake video being used, it becomes challenging for them to realize the deception. For example, in China, a scammer used face-swapping technology during video calls to steal $622,000 from a person's bank account.

Voice Messages

Today, it's effortless to clone someone's voice using AI. All that's needed is a three-second clip. Hackers use these deepfake voice messages to confuse or deceive people. This type of deepfake makes it challenging to distinguish between real and fake content. According to Forbes, in 2022, 37% of organizations fell victim to deepfake voice fraud.

Deepfake phishing has everyone concerned because it quickly becomes a growing threat. Thanks to AI, deepfake technology has become almost indistinguishable from reality. Initially, people trust it. According to reports, although surprising, incidents of deepfake phishing and fraud increased by three thousand percent in 2023.

Secondly, attackers can use deepfake to target specific individuals. They can determine their interests, hobbies, and social networks based on the victim's preferences. It's effortless for attackers to exploit the weaknesses of the target.

Thirdly, it's challenging to detect it. AI can mimic someone's writing style, clone voices almost flawlessly, create faces identical to humans, making it difficult to differentiate between real and fake. Consequently, identifying deepfake phishing has become increasingly difficult.

Ways to Detect Deepfake Videos

However, there are ways to detect fake videos using artificial intelligence-based tools. When in doubt about a video, carefully observe the nose, ears, feet, and hands of the people in the video. If there is inconsistency with the face, it is likely fake. Because AI cannot accurately recreate various body parts of humans. Additionally, if the voice doesn't match the lip movements, it indicates that the video is fake or created using artificial intelligence.