Use facial detection and recognition technology to prevent DeepFake
What is Deep Fake? The reason why an AI face scanner is necessary.
Deepfake is a combination of "deep learning" and "fake," a field of artificial intelligence, which is a technology that generates fake images or videos that are difficult to distinguish from the real world. The technology specifically imitates human characteristics such as faces and voices, making it look as if a real person has done or said something.
The origin of deepfake technology began when a member of Reddit, using the nickname "deepfakes," posted a pornographic video on the Internet that synthesized the face of a Gal Gadot actor. This led to the widespread use of the term deepfake, marking the beginning of an unprecedented era for digital content manipulation.
Impact of Deep Fake and ekyc solutions
With the rapid development of Generative AI, Deepfake has begun to make an impact in various industries.
Social media and entertainment
Create inappropriate images or videos by synthesizing celebrity faces into other people's bodies
It can be used for mockery or satirical purposes targeting public figures, which can lead to defamation.
Pornography
Infringement of honor and privacy with unsolicited pornographic content
Emotional and social damage caused by intimidation and coercion
Finance and Economy
Investor Deception for Fraud and Market Manipulation
Market Manipulation with Fake Announcement
Cybersecurity and Privacy
Social engineering attacks steal information
Voice/video manipulation for security system bypass
Politics and Elections
Public opinion manipulation with fake speech video of politician
Fake News Generation Causes Social Disruption
law enforcement and legal issues
Interrupt legal processes by manipulating video or audio records that can be used as evidence
Possible Deep Fake in eKYC
Selfie authentication is a very widely used method for identification in the course of eKYC.
This method is friendly from a user UX point of view and has a great advantage that it can be used on most devices. This is used to authenticate the real owner of identity information by comparing the picture of the ID with the face of Selfie.
In the past, it was difficult for scammers to cheat the platform while filming Seflie before Gen AI appeared, but now they can all too easily create real people like the example below and Deepfake.
Image used: Actor "Christian Bale"
Used AI Method : Image to image
Model : DreamShaper 8
Prompt : Use image reference at 75% similarity as “ethnic [variable]”, “age [variable]”
Production time: 1 minute
In addition to Deepfake, there are various types of fraud in the course of eKYC, such as the following.
All of these types of fraud can cause very serious problems in self-authentication, and there are cases where millions of dollars of remittances have already occurred through fraud.
Fraudsters Cloned Company Director’s Voice In $35 Million Heist, Police Find
“Audio and visual deep fakes represent the fascinating development of 21st century technology yet they are also potentially incredibly dangerous posing a huge threat to data, money and businesses,” - Jake Moore, a former police officer with the Dorset Police Department in the U.K.
Use facial detection and recognition technology to prevent DeepFake
ARGOS is offering a One-stop KYC solution. Not only does it provide accurate customer verification, but it also provides a variety of devices to stop scammers. Now let's see how ARGOS stops DeepFake.
Deepfakes using Generative AI are mostly transmitted to the 'digital environment'. After creating a fake video or image with Deepfake, you should try to authenticate it through the process below.
Upload and deliver the created image
Mark and deliver the created image through a display device or printout
1. Real Time base
Through Liveform, ARGOS provides a web browser-based screen that allows customers to launch in just two minutes without any development process. It is already well known that technologies such as Deepfake are the most used on web browsers. In particular, it spreads more rapidly in free file upload environments such as social media.
ARGOS provides an option to block the upload function when providing live form and to set it to be used only by real-time camera shooting.
The 'Liveform live camera recording only' feature allows customers to block upload-based authentication, preventing the delivery of created images or images, providing a safer authentication experience.
Setting screen in the dashboard of ARGOS
Liveform live camera recording only
ARGOS Liveform Progress Screen
Setting the option will allow the browser camera or Native camera to only shoot in real time.
2. Passive Liveness
Now that the path to upload has been blocked, the scammer must attempt to scam the image or video created in a separate 'displayed form' such as a display device.
If the ARGOS system requires only real-time camera shots, the scammer will attempt to authenticate through pre-made images on a tablet or other mobile display.
Typically, the solution providers offer Active Liveness, which requires randomly assigned behaviors such as facial nodding, blinking, and opening your mouth, and recognizes these behaviors to confirm Liveness.
Core of Active Liveness
Pre-generated images or images will not be able to follow the required actions in real time.
Examples of what might happen if you're doing active living with your facial expression
The action that requires the specified action must require additional time for the user to complete the step, and the 'pattern of action' required is indicated, exposing a vulnerability that reverse engineering can prepare in advance.
ARGOS always values the 'user's position' - identity verification service providers
Scammers like Deepfake pose a big threat, but they should not inconvenience normal users to prevent them.
Considering the business, a simple, easy, and fast process is good for the user. This is why ARGOS provides Passive Liveness. Complex processes are directly linked to user churn.
ARGOS provides customers with Passive Lives at any time through options as follows. It is a Level 2 product certified based on ISO 30107-3, and boasts the highest accuracy in the world.
Key to Passive Lives
You should be able to spot scammers without the process of a normal user changing.
The advantages of ARGOS' Passive Lives are as follows.
1️⃣ Fast
It takes less than a second. A unique single image approach uses the same selfies taken for facial recognition to judge liveliness.
2️⃣ Frictionless
A single image approach doesn't require engagement at all for a better user experience increases the rate of automation and reduces customer churn.
3️⃣ Accurate
Complying with iBeta/NIST Level 1 and Level 2 PAD, we received a perfect score. The software performs millions of active health checks each month.
4️⃣ Works for Everyone, Everywhere
We are investing in creating algorithms that provide unbiased results across demographics such as race, gender, and age.
5️⃣ Cost Efficient
Unlike other solutions that require high bandwidth internet connectivity and large server deployment to handle video, infrastructure costs are reduced.
Below is a comparison table of Passive and Active lives.
Category | Passive liveness | Active liveness |
|---|---|---|
User interaction | Operate without user action | Requirese direct user participation (e.g look up, nodding, blinking) |
User awareness. | Unrecognizable during liveness. | Recognizable during liveness. |
Algorithm | Facial features or the texture of the captured image. | Analyze behavior |
Process | No additional processes. Concurrently happening in the backend in the process of taking a selfie. | Requires additional layers to be installed to require behavior and measure behavior. |
ARGOS' 2023 fraud attempt detected through Liveness detection selfie
Even before the Deepfake trend, ARGOS detected the importance of Liveness and has continuously provided Liveness services to customers. Below is a trend of detection of Passive Liveness detected by ARGOS in 2023.
Considering that those who attempt to commit fraud proceed with multiple attempts rather than one attempt, it can be seen that there are as many as 1% of the total attempts.
ARGOS continues to work to detect these fraud attempts and provide safe service to customers.
ARGOS' Efforts for best identity verification Solution
In addition to features to prevent Deepfake, ARGOS is equipped with a variety of features to make it available to customers with One-stop solutions such as accurate eKYC and fraud prevention.
At the link below, you can see the various functions of eKYC provided by ARGOS.
eKYC's journey with ARGOS
Did you enjoy using ARGOS eKYC for free? If you need more eKYC verifications or have any questions about the service, please don't hesitate to contact us.
ARGOS will continue to strive to make your onboarding journey more convenient and secure. Join us in making your online world safer and more convenient!