Sumsub Reports 180% Rise in Sophisticated Fraud Globally

Fraudsters are executing fewer attacks but achieving significantly higher success rates through coordinated, multi-step techniques, according to Sumsub's latest research tracking more than 4,000,000 fraud attempts worldwide.

The verification firm's Identity Fraud Report 2025-2026 reveals a fundamental transformation in how criminals operate. Sophisticated fraud attacks, which combine several advanced techniques within a single verification attempt, have surged 180% globally during 2025. 

This marks a decisive break from traditional high-volume, low-success approaches that dominated the fraud landscape until recently.

The shift has particular implications for financial services and fintech firms that rely on digital onboarding. 

Rather than flooding systems with thousands of crude attempts, criminals are deploying crafted attacks designed to slip past automated defences and manual review processes alike.

AI tools fuel realistic document forgery

The proliferation of consumer-facing artificial intelligence platforms has equipped fraudsters with tools previously unavailable outside specialist communities. ChatGPT and Gemini now account for 2% of all falsified documents Sumsub processes globally, with the firm projecting double-digit growth in AI-generated document fraud throughout 2026.

Deepfake attacks have accelerated sharply across European markets in particular. The UK recorded a 94% increase in deepfake fraud attempts during 2025, trailing only France at 96%. Spain saw an 84% rise whilst Germany experienced a 53% increase.

One in five fraudulent verification attempts in Europe this year involved an edited or forged identity document, and as a result, some 72% of EU companies now anticipate more attacks using artificial intelligence, particularly deepfakes and synthetic identity documents.

Pavel Goldman-Kalaydin, Head of AI/ML at Sumsub, says legacy fraud prevention metrics no longer reflect operational reality. “The traditional measures of success in fraud prevention are quickly losing meaning,” he says. 

“The threat has shifted from quantity to quality, and resilience now depends on how quickly organisations can detect anomalies, analyse behavioural data and adapt their defences to emerging threats in real time.”

European firms report widespread financial losses

The impact extends beyond technical concerns to financial performance. Some 64% of European companies reported fraud-related financial losses during the period studied, while 36% cited reputational damage. More than half of European businesses became fraud victims in 2025, despite total fraud rates declining marginally by 0.4%.

This apparent contradiction stems from the efficiency gains criminals have achieved. Attackers no longer need multiple attempts to penetrate verification systems, making detection more difficult for organisations relying on pattern recognition and volume-based alerts.

The professional services sector experienced the sharpest increase, with identity fraud attempts rising 232% globally during 2025. Notably, legal, consulting and accounting firms have become targets due to their access to sensitive client data and continued dependence on manual onboarding processes.

Online media platforms recorded a 6.3% fraud rate in the UK, while the internet dating sector reached the same level. Moreover, romance scams are increasingly using AI-generated personas and deepfake videos to establish credibility with victims.

Manual verification creates vulnerability

The research identifies a structural weakness in how European businesses approach fraud prevention. Despite operating within established digital identity programmes and comprehensive regulatory frameworks, 37% of European firms continue to rely on manual verification processes.

This creates exploitable gaps for criminals deploying sophisticated AI-enhanced techniques. Manual reviews struggle to identify subtle inconsistencies in synthetic documents or detect coordinated attacks that span multiple channels and identities.

Jacob Thompson, VP of Sales at Sumsub, says the disconnect between awareness and implementation remains acute. “Nearly 3 in 5 European consumers were victims of fraud in 2025,” he says. “For now, the most visible shift is not in how much fraud exists, but how efficiently it is executed.”

The report also highlights consumer vulnerability beyond business systems. Some 87% of Europeans cannot identify money muling schemes, leaving them susceptible to recruitment for transferring illegally obtained funds.

Autonomous agents represent next threat

Sumsub anticipates fraud rates will stabilise or decrease during 2026, but predicts greater adoption of autonomous agents capable of conducting sophisticated, end-to-end attacks with minimal human oversight.

This industrialisation of fraud will make cross-channel manipulation more accessible while enabling criminals to deploy multiple coordinated synthetic identities simultaneously. The convergence threatens to overwhelm verification systems that operate through isolated checks rather than continuous monitoring.

Pavel says the evolution demands corresponding advances in defensive capabilities. 

“Attackers gain deepfakes, synthetic IDs and autonomous fraud agents; defenders gain behavior modeling, millisecond anomaly detection, and self-learning systems,” he says. 

“The next frontier is verifying AI agents themselves - confirming not just who you are, but who acts on your behalf.”