Sumsub launches adaptive deepfake detection system to strengthen digital fraud prevention measures

by

Rei Fortes

-

5 hours ago

Sumsub launches adaptive deepfake detection system to strengthen digital fraud prevention measures

Singapore – Sumsub has introduced ‘Adaptive Deepfake Detector’, an updated deepfake detection system designed to improve the identification of increasingly sophisticated online fraud. The new technology uses continuous machine learning updates in real time, aiming to address weaknesses found in conventional systems that rely on periodic offline model refreshes.

Traditional fraud detection models are typically updated at set intervals, which can leave gaps lasting weeks or months before new threats are recognised. During that time, emerging scams may evade controls and affect both businesses and users. 

The launch reflects growing concern over the speed and complexity of digital fraud. Multi-stage fraud attempts rose by 180% in 2025 and accounted for 28% of all fraud cases identified on its platform worldwide. The use of AI-generated deepfakes has also continued to increase since 2023 across several markets.

“Modern deepfakes can no longer be detected by the human eye, and decision-making should be based on multiple signal analyses in real time,” Nikita Marshalkin, Head of Machine Learning at Sumsub, stated. 

“That’s why we launched our upgraded Deepfake Detector, offering clients not just a tool, but rather an online learning system that combines advanced document checks, device intelligence, and fraudulent networks analysis to complement deepfake detection capabilities.”

The Adaptive Deepfake Detector is designed to analyse information from a range of sources rather than relying on a single anomaly indicator. These include identity documents, geolocation, IP data, device characteristics, facial liveness checks and patterns that may indicate linked fraudulent accounts or networks.

Additionally, the system continuously monitors new activities, removing the need for scheduled retraining cycles or extensive manual reviews. It dynamically adjusts detection thresholds in response to changing threats in an effort to maintain high levels of accuracy.

The development highlights a broader shift in fraud prevention towards adaptive and automated risk management tools as businesses face faster-moving AI-enabled threats.

Recognise the innovators redefining commerce at the Retail & E-commerce Excellence Awards Asia Pacific 2026! Taking place this December 2026, we celebrate the region’s most impactful retail strategies, standout e-commerce experiences, and forward-thinking leaders—submit your entries today!
Honour the women shaping the future of marketing and technology at the Empowered Women Awards 2026! This December 2026, we celebrate inspiring leaders, changemakers, and rising voices driving impact across the industry—submit your entries today!
Share

RECENT ARTICLES

Sumsub launches adaptive deepfake detection system to strengthen digital fraud prevention measures
Workato expands Philippines presence to support enterprise AI growth
Philippines digital service outage highlights strain on government systems
Singtel Singapore launches AI programme to support SME digital transformation
ASEAN Foundation launches regional programme to counter online scams in Southeast Asia
Ellipse 3

RELATED ARTICLES

ASEAN Foundation launches regional programme to counter online scams in Southeast Asia
Sumsub taps Go Digital Philippines to strengthen digital trust, AI governance across ASEAN
1
Ellipse 3

FEATURED ARTICLES

UpTech Media, MARKETECH APAC unveils second edition of ‘Retail & E-Commerce Innovation Summit’ 2026 in Malaysia
‘Retail & E-Commerce Excellence Awards Asia Pacific
Empowered Women Awards 2026 to honour leading women trailblazing the technology and marketing industries

Subscribe to UpTech Media Newsletter