Online identity fraud continues to be a major problem across industries, with an estimated 1 in 25 verification attempts now involving attempts to impersonate another person. The problem is particularly acute in financial services, where online fraud is used to open accounts, take out loans, gain access to credit, and more. During 2025, fraudulent attempts in online financial services exceeded 5.5 percent of all verification attempts, representing a significant increase from 2024’s figures.
Compared to other types of fraud, research shows the growth rate for traditional document fraud presentation – i.e., when a document’s data has been physically altered, or a counterfeit document has been fabricated from scratch – has been relatively stagnant the last few years.
Some fraud prevention professionals may view this as a positive sign that fraudsters are realizing that online identity verification systems have become “too good” at discerning and rejecting fake and altered IDs, and have moved on to more sophisticated tactics.
However, now is definitely not the time to lower your guard or be lulled into a false sense of security that document fraud is a lesser concern. We expect that synthetic identity fraud will drive the next big evolution and uptick in document fraud.
Synthetic Identity Fraud Demands a Refocus on Documents
Synthetic identity fraud occurs when fraudsters blend real and fake data to create documentation that’s ‘good enough’ to fool most online identity verification checks. Today, a whopping 80 percent of new account fraud can be attributed to synthetic identities.
The rise in Fraud-as-a-Service (FaaS) means just about anyone can access sophisticated AI tools, such as face-swapping (a form of deepfakes) and face document creation (using ID templates widely available on the dark web). For instance, today, a fraudster can simply take a picture of someone else’s ID, upload it to their computer, swap in the fraudster’s own face, print it onto a PVC card, present it – and they’re off to the races. A new account can be created, a loan taken out, or a credit line created.
Unfortunately, synthetic identity fraud is currently costing businesses billions, and it’s likely only going to get worse. Estimated losses from synthetic identity fraud are expected to hit at least $23 billion annually in the U.S. alone by 2030. And it’s hitting all industries, from retail and e-commerce, to automotive lending, to property lending, and government/public benefits. Currently, 44 percent of organizations across industries overall rank synthetic identity fraud as the top-tracked fraud type.
Synthetic identity fraud often entails fraudsters combining a stolen SSN (often from deceased individuals, or children, or the elderly who don’t closely monitor their credit) with fake names, addresses, and phone numbers to create a new, fake “Frankenstein” identity. Fraudsters then make small purchases over months or years, establishing a credible credit history. Once high credit limits are established, the fraudster then makes purchases of high-value, easy-to-resell goods (for instance, electronics and luxury items in retail) in a process known as “busting out.” Once the retailer becomes aware of the fraud, it’s usually too late – the fraudster has disappeared, and the business is left with no real person to hold accountable.
How AI Makes Online Identity Verification More Complex
In the days before AI, businesses typically verified identity by cross-checking the face on a document with the person presenting it, to make sure they matched. But today, AI-enabled trickery such as synthetic identity fraud means that traditional biometric face matching is no longer sufficient to authenticate an individual. Face matching will show that the person on the document is the person presenting, but this does not provide any guarantees whatsoever that the person is actually truly associated with the document or its details.
The world of synthetic identity fraud requires a renewed focus on document verification as part of a comprehensive, multi-layered approach – verifying document legitimacy; biometric checking (i.e., face matching to ensure the face on the document matches the face of the individual presenting); analyzing behavioral biometrics (such as keystroke dynamics and device movements); and assessing technical signals like device and network fingerprinting. Integrating these continuous, non-intrusive authentication methods significantly strengthens fraud detection by revealing anomalies that static checks often miss.
The benefits of such an approach are many, including:
- Reducing Fraud Loss: Across industries, approximately one in twelve newly created accounts is estimated to be fraudulent. Stopping synthetic identities and corresponding bogus IDs at the point of entry enables businesses to nip fraud-related losses in the bud. Businesses looking to expand globally will need document verification capabilities that deliver superior accuracy across a wide variety of document specimens (driver’s licenses, passports, and more) and key countries and geographies.
- Solving the “Friction Versus Security” Paradox: Slow, inconvenient online onboarding processes have long plagued businesses, particularly financial services firms, with one study reporting user abandonment rates as high as 50 percent when digital account opening takes 3-5 minutes or more. New approaches are both extremely fast and reliable, helping maximize conversions.
- Operational Efficiency: Near-perfect accuracy and automation enable businesses to scale globally without a corresponding increase in manual review costs or headcount. Moreover, maintaining “human-in-the-loop” capabilities enables human oversight of highly atypical specimens, ensuring a balance between optimal security and a stellar user experience for the masses.
- Regulatory Compliance: Businesses can also benefit from cross-industry recognition of fraudulent patterns (for instance, the same synthetic profile and supporting documentation used across multiple e-commerce platforms to abuse return policies). Financial services firms, in particular, can benefit from the ability to cross-reference users against global sanctions lists and government registries, helping to ensure both KYC and AML compliance.
Given the rise of AI, synthetic identity fraud detection – including an upleveled ability to detect fake and altered documents – is taking on a new level of urgency. While more traditional forms of document tampering and forgery may be slowing down, AI-based methods are accelerating, and now is not the time to relax the focus on documents. This will be particularly important as users increasingly expect all online platforms to be safe, secure, and defensive.
About the Author
Iryna Bondar-Mucci is a Fraud Platform Lead at Veriff. She has a master’s degree in cybersecurity and steers Veriff’s overall fraud posture, leads scalable defense initiatives, and is also one of the authors of the annual Veriff Identity Fraud report. Ira serves as strategic lead for fraud mitigation across Veriff.
Iryna can be reached online via LinkedIn here and at the Veriff website
