Sumsub Launches Adaptive Deepfake Detection Solution to Combat Rising Digital Fraud

• •Sumsub's Adaptive Deepfake Detector adapts to new threats within hours using machine learning.
• •Multi-step attacks increased by 180% in 2025, highlighting the urgency for advanced detection tools.
• •Traditional deepfake detection methods are inadequate as modern deepfakes evade human detection.

On April 30, 2026, Sumsub introduced its Adaptive Deepfake Detector, a new solution designed to address the limitations of traditional offline deepfake detection methods. This model utilizes machine learning-driven detection with instant online self-learning updates, allowing it to adapt to emerging fraud techniques within hours. The urgency for such technology has escalated, as multi-step attacks increased by 180% in 2025, comprising 28% of all fraud detected on the Sumsub platform. The rise of AI-generated deepfakes since 2023 has posed significant challenges for risk management teams, as modern deepfakes are undetectable by the human eye. Sumsub's tool combines advanced document checks, device intelligence, and analysis of fraudulent networks to enhance detection capabilities. The solution aims to provide a comprehensive approach to counteract the growing sophistication of online fraud. Current detection methods relying solely on visual content are insufficient, necessitating a broader context analysis of user sessions. The launch signifies a critical development in the ongoing battle against digital fraud.