Full-cycle verification platform Sumsub has launched a Deepfake Detection feature built into its Video Identification solution.
Sumsub’s Deepfake Detection feature represents a major leap forward in the fight against AI-based identity fraud, particularly as fraud tactics become increasingly sophisticated.
In the last year, Sumsub has performed over one million video authentications for companies, securely verifying their customers’ end users. Along the way, Sumsub’s AI and ML team noticed a pattern in video verification interviews that were prone to deepfake attacks, and therefore decided to augment their existing video recognition solution with Deepfake Detection.
This follows the release of Sumsub’s enhanced Deepfake Detection solution, which is integrated into the Liveness in-house product.
Unlike methods that focus on detecting deep fakes in still images or recorded videos, Sumsub’s new solution works in real time during video interviews, further ensuring the identification process. Enhanced deepfake detection technology is designed to detect and block fraudsters who attempt to manipulate real-time video interviews for malicious purposes.
Video ID verification is mandatory in several jurisdictions, including Germany, Switzerland, Austria and Estonia. However, it can also be used as a reliable last resort for user verification in most markets when businesses want to provide an extra layer of security during onboarding.
“AI technology and its risks are not new and we have spent many years developing deepfake detection tools long before they were introduced into the mainstream. As AI technologies advance, so do the tools available to fraudsters. Unfortunately, cases like these will continue to happen, making our first-of-its-kind solution that integrates deep real-time video misidentification detection so imperative for business and consumer security,” said Andrew Novoselsky, Chief Product Officer on Sumsub.
“We often see increasing demand from our customers to conduct video interviews, so we understand the importance of ensuring that these are completely secure. Video recognition is often seen as the ultimate defense against embedding fraud, but as deepfake technology evolves, trusting our eyes during live video is no longer foolproof. To address this, we are actively enhancing our detection capabilities. As deepfakes continue to impact business operations as well as everyday people, we are committed to continuously working to expand our detection capabilities to protect companies and their customers,” added Pavel Goldman-Kalaydin, Head of AI & ML at Sumsub .