Sumsub Research Finds Digital Fraud Networks With Hundreds of Members

One out of every hundred users of a digital platform or service is a member of a fraud ring, according to new research from Sumsub.

That’s one finding from the company’s internal review of anonymized user verification data from 2023. The information was collected from a client base spanning financial services, crypto, online gaming, digital marketplaces, and dating platforms, with Sumsub’s anti-fraud platform having analyzed more than 150,000 fraud attempts on a monthly basis.

Sumsub is now sharing some interesting findings based on that research, including the wide variability in the sizes of different fraud networks, which can range from a group of three individuals to a networks in the hundreds. But the concerted actions of so many individuals can itself raise red flags for anti-fraud systems.

“In one case, a group of several dozen crypto exchange applicants in Estonia raised suspicion by uploading identical Proof of Address documents from an unlicensed foreign bank,” said Sumsub’s Head of AI/ML, Pavel Goldman-Kalaydin. “This revealed potential attempts to issue many crypto cards to the same address. This is just one case of how serial fraud operates, other instances include money muling schemes, tech support scams, ransomware and phishing attacks, and account takeovers.”

Sumsub’s recently upgraded Fraud Prevention Solution encompasses a number of tools and techniques aimed at automatically discovering potential fraud threats. “To flag suspicious user activity, our system uses Identity Verification, Behavioural Intelligence, Device Fingerprinting, Fraud Risk Scoring, Deepfake Detection, Email and Phone Risk Assessment, and AI-based Event Monitoring,” explained co-founder and CTO Vyacheslav Zholudev, adding, “The solution now also applies Fraud Network Detection to stop serial fraud.”

The company’s deepfake detection technology in particular has garnered considerable attention in the biometrics space over the past year. In June of 2023, Sumsub announced an enhanced version of its liveness detection solution that would use 3D face mapping to flag potential deepfakes; and in October it launched a set of Machine Learning-driven AI models on an open source basis, inviting the AI research community to leverage them in innovative systems for deepfake detection.

–

March 26, 2024 – by Alex Perala