Remote Digital Onboarding! Are you being spoofed?

There are many different approaches to customer digital onboarding, from video KYC to automated enrollment methods involving ID and selfie capture. Of course, online service providers are all trying to offer a frictionless user experience.

Frictionless? This word is overused and is often not an accurate description of the whole process. One way to measure frictionless is how long it takes to onboard a new customer. The average time across many platforms can be anything from 15 to 30 minutes. It can also be much longer if a user must undergo an online video interview to prove their identity and whether they are a real person.

Biometrics help deliver a seamless user experience, but there is a bigger problem. Spoofs are common! Spoofing is when a fraudster tries to defeat the biometric system with photos or video replays. However, some systems are not robust enough to detect a potential spoofing attack. Believe it or not, it is quite easy to fool biometrics with photos, masks and deepfake videos. So while biometrics is used to verify a user’s identity, liveness detection is needed to ensure that they are 100% real. Fraud is still very much on the rise and fraudsters are a lot more sophisticated in understanding how to bypass some onboarding methods. Liveness detection is now considered an essential part of the onboarding process.

With the explosion of cryptocurrencies, many start-up crypto providers are being exploited by fraudsters because their KYC processes are too weak. Detecting criminals is a serious challenge.

Therefore robust anti-spoofing methods are needed. But how do you spot a spoof? Here is an excellent article from Sumsub on how fraudsters are able to potentially bypass some unsophisticated liveness methods: https://sumsub.com/blog/learn-how-fraudsters-can-bypass-your-facial-biometrics/. The article explores both spoofing and bypassing, so it is worth a read.

In order to combat spoofs, some KYC and onboarding operators have had to introduce liveness checks into the onboarding process as an extra security step. We refer to these as active liveness methods. Either making a short video recording or users making certain random gestures into a camera. These liveness checks can vary with different tasks and of course are time consuming. Yet, many providers claim they are frictionless when studies have shown that there are clear indications of abandonment by new customers. When a new customer is challenged to prove they are real, the criteria is often difficult to comprehend and therefore not properly executed. In a recent Tier 1 bank study, it had shown there was up to 50% abandonment in the customer journey at the point of where the active liveness check took place, because after three attempts the customer was either rejected or had given up!

Also, in some countries, enterprises and operators have implemented video onboarding which is an even longer process for new customers to go through. Essentially, it's an online interview where your personal details and ID information are obtained. The process involves a huge facilitation cost for any operator and it can also be frustrating for new customers who have wait in line for a customer service agent to attend to them before the process can begin. If you have ever tried to switch a Utility provider online, you will know what I mean! The process is extremely long-winded particularly going through all the security checks and providing personal information. The term we make it easy to switch is almost a joke!

Therefore, claims about being frictionless are sometimes wildly exaggerated. This is where a new approach is needed. 

AI technology has rapidly advanced to new levels and as such, so have liveness technologies. Static or passive liveness is the new generation of liveness detection systems. AI neural networks are being used to determine whether a person is real from a single selfie. The same selfie used for ID matching can be used for liveness checking thus removing the extra step for an active challenge. Furthermore, it would be difficult for any fraudster to determine whether there is actually a liveness check taking place. How would they know?

Some operators and regulators find it difficult to understand how a single static image can prove liveness but today it has been achieved with AI, so you can read more about the technology here.

Static liveness checks have a clear advantage over any active challenges, as there is no need for customers to be troubled with nodding their head or moving a device around to prove that it is a real person. These kinds of outdated challenges create for a very poor user experience. Security is paramount but so too is providing a robust seamless process for ensuring spoofs are detected without any hassle for end users.

ID R&D's passive facial liveness detection is being used by over 60 global IDV, KYC and onboarding tech partners and processing millions of liveness checks every month. You can read more about how ID R&D's liveness solution here: https://www.idrnd.ai/, or see a video explainer here. If you would like to try a liveness demo yourself, please contact ID R&D: [email protected].