Cybercriminals on the dark web use deepfakes to generate AI fake videos and images for illegal activities, including fraud, blackmail, stealing confidential data, stock manipulation, sextortion, and other crimes.
In this post, we will let you know all the missuses of deepfakes technology that are used on dark web marketplaces. But let us first know about deepfakes, how they work, and the technologies that are used in deepfakes.
| Key Takeaways Deepfakes use GANs: Generative adversarial networks, NLP: Natural language processing, and CNNs: Convolutional neural networks, high-performance computing, and autoencoder AIs to accomplish their work. Deepfakes are synthetic media files, including imagery, video, or audio, that manipulate and replace another person’s face or voice. The developers of deepfakes on dark web charges from $300 to $20000. However, the charges are based on the complexity of the work. Deepfakes can be easily recognized as fake by analyzing the reflection of light in the eyes, blurring around the ends of the face, lack of blinking and the irregularities in the hair, mood patterns, scars, etc. |
An Overview of Deepfakes
The term Deepfakes combines with two words deep learning and fake. Deepfakes are basically synthetic media files, including imagery, video, or audio. It typically features a specific individual who manipulates and replaces another person’s face or voice.
However, this work is accomplished using generative AI: artificial intelligence-powered neural networks, also known as GANs: generative adversarial networks. It processes data, creates patterns, and learns much like the human brain does.
The deepfake technology was developed at the beginning of the 1990s by researchers at academic institutions. Industries recently adopted this technology for adding visual effects and animation. Right now, the extensive availability of computer science technology and the growing accessibility of AI enable virtually anyone to make highly realistic fake content.
In fact, the number of deepfake videos is growing by 900% a year. The manipulation of content to influence audiences is not new. But the line between what is real and what is fake has become razor-thin.
How Deepfakes Work?
Deepfakes basically use 2 algorithms to create and refine fake content.
- Generator: It built a training data set based on the desired output for creating the initial fake content.
- Discriminator: It analyzes how realistic or fake the initial version of the content is.
However, the procedure is repeated to let the generator progress at producing realistic content. At the same time, the discriminator becomes more skilled at spotting flaws for the generator to be accurate. The amalgamation of generator and discrimination algorithms makes a generative adversarial network.
GAN uses deep learning to recognize patterns in the real image and then use those patterns to produce the fake image. The GAN system views the image and video from various angles to capture all the details and perspectives like behavior, movement, and speech patterns when producing a video and an image.
Different Approaches for Creating Deepfakes
Here, we have shared some specific approaches for creating deepfakes.
- Audio Deepfakes: GAN duplicates the audio of a person’s voice and creates a model based on the vocal patterns. Then, uses the model to take the voice and say anything the creators want for deepfake audio with the help of NLP. The video game developers basically use this approach.
- Video Deepfakes: A neural network-based deepfake autoencoder analyzes the content to get relevant attributes of the target, like facial expressions and body language, for creating a video. Then, it executes these features onto the original video.
- Lip Syncing: The recurrent neural networks map a voice recording to the video, making it look like the person in the video is speaking the words in the recording in lip-syncing deepfake.
Computing Technologies Required to Create Deepfakes
Because of the following technologies, creating deepfakes is becoming easier, more precise, and more relevant.
- GANs: Generative adversarial network technology is used to develop the deepfake content using the generator and discriminator algorithm.
- NLP: Natural language processing is used to produce deepfake audio. The algorithm of NLP technology analyzes the attributes of a target speech and then produces original text via those attributes.
- CNN: Convolutional neural networks analyze patterns in visual data that are used for facial recognition and movement tracking.
- High-performance Computing: This technology offers the necessary power required to generate deepfakes.
- Autoencoders: This technology detects the relevant features of a target, like facial expressions and body movements, and then executes these features onto the source video.
The Missuses of Deepfakes Technology on the Dark Web
Deepfake technology has some positive points, as it is used in generating music videos, caller-responsive services, and customer support. In addition to these uses, the negative points easily overwhelm our society. Below, we have shared the misuse in the dark web world of deepfake technology.
1. Spreading Fake News
Currently, fake news is disruptive because of its prevalence and influence over the public. Deep fake tech helps its users to make more credible hoax material than ever, like fake speeches and interviews involving celebrities, politicians, and influential figures to spread misinformation and fake news to manipulate the public.
2. Cyber Attacks
New technology always brings about new solutions — and that includes crime. The recent trend in deepfakes could bring online criminality to a new level. Consider that this technology will become more available as time advances. So, it will become a standard AI tool for social engineering (which is already one of the most effective cybercriminal techniques). Social engineering relies on human emotions as it is the weakest link in the digital security chain for manipulating a person’s behavior to make them do something they would usually avoid.
3. Sextortion
This technology is also used to create fake adult and sexually explicit material featuring the victim person without their consent, violating their privacy and dignity. This scam is known as sextortion, which is a common thing in dark web forums. Because of deepfaking, it has become very easy to replace and change a face and voice to another one in a video. Dreadful but accurate, you would never spot a single difference till you are told. It is so easy that any video alteration is possible, including yours and me.
4. Phishing Attacks
Deepfakers can use this technology for a new brand of phishing attacks. A phisher can mimic someone else to encourage new victims into action. For instance, a video could be generated with a CEO announcing that a company has lost all customer data. However, everything can be on point with the deepfaked voice or video. An attacker could sell the data of the company on the dark web and deep web marketplaces, and on the contrary, the victim could lose the company.
5. Blackmailing and Bullying
The common example of blackmailing and bullying is when a target image is put in an illegal, inappropriate, or otherwise compromising situation, such as lying to the public, engaging in explicit sexual acts, or taking drugs with the help of deepfaking technology. These videos are used to extract a victim, collapse a person’s reputation, get revenge, or simply cyberbully them. The most usual blackmail or revenge use is nonconsensual deepfake adult content, also known as revenge porn.
6. Fraud
Deepfakes are also used to impersonate or defame an individual, organization, or brand to obtain identifiable information such as bank accounts and credit card numbers. This can sometimes include impersonating executives of companies or other employees with credentials to access sensitive information and creating fake reviews and testimonials, which is a major cybersecurity threat.
7. False Evidence
Deepfakes involve the assembly of false images or audio that can be used as evidence of implying guilt or innocence. In a legal case, these deepfakes are so good that even the prosecution may not be able to decipher the real ones from them. It led them, in many cases, to try introducing them as evidence at trial. And because of their genuine form, it can be very difficult for defense lawyers to verify that the evidence is dishonest.
8. Stock Manipulation
Deepfakes materials are used to affect a company’s stock price. For example, a fake video of a chief executed officer making damaging statements about the company could lower its stock price. A fake video about a high-tech advance or product launch could increase a company’s stock. Furthermore, they could release a deepfake video of the organization’s CEO making a positive statement about technological inventions to progress the company’s stock. They could also release a video with damaging content to lower the company’s stock price.
9. Fake Identity
There are thousands of pseudo-accounts created using photographs from deepfakes. Anonymous users create deepfake non-existent pictures with prominent and use fictional names with the intention of hiding their identity. The users then actively engage in forums and discussions where they incite people with propaganda. Others use the accounts to dupe unsuspecting users to lure them into online scams.
The Danger of Deepfakes on the Dark Web
According to reports, deepfakes employee burnout and ransomware attacks from the dark web. The forums of the dark web discuss deepfakes generating its services, methods, lessons, and news of evolving tactics and technology. These discussions are largely concentrated in English- and Russian-language criminal forums. But related topics were also observed on Turkish-, Spanish- and Chinese-language forums.
However, the most common deepfake-related topics on dark web forum services are editing videos and pictures, how-to methods and instructions, requirements for best practices, sharing free software downloads and photo generators, general interests, and announcements on advancements in deepfake technologies. Moreover, there is a strong surface web presence and interest in deepfake technology, consisting of open-source deepfake tools, dedicated forums, and discussions on popular messenger applications such as Telegram and Discord.”
Here, we have shared some cyberthreat activities using the deepfake that we found on the dark web.
- A threat actor offered services on hack forums of the dark web for $20 per minute of fraudulent video with realism.
- The developers of deepfakes on dark web charges from $300 to $20000 based on the complexity of the work.
- Another threat actor posted their willingness to pay $16000 for deepfake services, including video and photo editing.
- An anonymous researcher stated that almost 244625 videos had been uploaded to the top adult websites of the dark web that were set up from deepfake technology.
- The survey of 125 cybersecurity incidents stated that the security incidents involving deepfake use in the dark web have risen in the last year.
- Criminals from the dark web often request deepfakes for cryptocurrency scams, cracking online accounts, or adult content.
Future of Deepfakes
Currently, deepfakes technology is in its infancy, but it lets the cybercriminals on the dark web harm others severely. It can be easily recognized as fake; here are a few examples of recognizing the deepfakes.
- The reflection of light in the eyes
- Blurring around the ends of the face
- Lack of blinking in eyes
- Irregularities in the hair, mood patterns, scars, etc.
However, machine learning and artificial intelligence technology are quickly maturing and increasingly becoming more difficult to detect. That is why the imperfections will disappear over time as the software used to produce deepfakes improvises hastily.
There are many wits from technology companies trying to combat deepfakes. It will be a constant struggle till companies finally outplace deepfake creators who, more often than not, can quickly find new ways to stay ahead of detection methods and continue to cover their tracks.
Summing UP
The dangers of deepfakes on the dark web lie in the potential to get misused for malicious purposes. That can include spreading fake news, cyberattacks, sextortion, fake identity, stock manipulation, blackmailing, bullying, and other cybercrimes. Furthermore, bad actors can use this technology to make fake videos of people saying or doing things they never actually said or did.
The dark web world is filled with service methods, lessons, and news of evolving tactics and technology for generating deepfakes and charges a huge amount for services. Thousands of people use these services for such cyber-attacks.