In 2024, digital document forgery overtook physical counterfeiting for the first time, deepfake attempts hit one every five minutes, and US cybercrime losses reached a record $16.6 billion. Document fraud has crossed a threshold: AI has made forging a convincing document cheap, fast, and scalable, and the data now shows it.
This page gathers the most current, source-backed statistics on document and identity fraud worldwide — each attributed to its publisher and period — so you can cite the numbers with confidence. We separate observed data from projections throughout, and update the page as new reports land.
Key statistics at a glance
The most-cited global figures on document and identity fraud, each attributed to its publisher and period.
| Metric | Figure | Source (period) |
|---|
| Digital vs physical document forgery | 57% digital / 43% physical — digital overtakes physical for the first time | Entrust (2024) |
|---|
| Digital document forgeries, year on year | +244% (and +1,600% since 2021) | Entrust (2023 to 2024) |
|---|
| Deepfake attempt frequency | One every 5 minutes | Entrust / Onfido (2024) |
|---|
| Deepfake fraud growth | +700% | Sumsub (Q1 2024 to Q1 2025) |
|---|
| Sophisticated identity fraud | +180% year on year | Sumsub (2025) |
|---|
| US cybercrime losses (all types) | $16.6 billion (+33% YoY) | FBI IC3 (2024) |
|---|
| Business Email Compromise losses | $2.77 billion across 21,442 incidents | FBI IC3 (2024) |
|---|
| Largest single deepfake fraud | about $25.6 million over 15 transfers (Arup) | reported Feb 2024 |
|---|
| GenAI fraud losses, US (projection) | $12.3bn (2023) to $40bn by 2027 | Deloitte (projection) |
|---|
| Nigeria financial-sector fraud losses | ₦52.26 billion | NIBSS (2024) |
|---|
| Africa document fraud | +125% in H2 2024 | Smile ID |
|---|
AI and digital document forgery
The single most important shift in 2024 was structural. According to Entrust's 2025 Identity Fraud Report, digital document forgery overtook physical counterfeiting for the first time, accounting for 57% of all document fraud versus 43% physical — a reversal of the historic pattern. Digital document forgeries rose 244% year on year and 1,600% since 2021, when almost all fraudulent documents were physical counterfeits. National ID cards were the single most-attacked document type, at 40.8% of global document attacks.
The driver is AI. Generating a convincing fake — a doctored statement, a synthetic ID, a fabricated certificate — once took skill and time; in 2026 it takes a prompt. That is why losses can rise even as the number of attacks plateaus: each attack is cheaper to produce and harder to detect.
Deepfakes
Deepfakes moved from novelty to production-grade fraud tool across 2024 and 2025. Sumsub's Identity Fraud Report 2025 to 2026 recorded a 700% increase in deepfake fraud between Q1 2024 and Q1 2025. Entrust and Onfido clocked a deepfake attempt every five minutes in 2024, and reported deepfakes accounting for roughly one in five biometric fraud attempts.
The Arup case, reported in February 2024, remains the highest-profile single incident: a finance employee was deceived by a deepfake video conference impersonating executives into making 15 transfers totalling around $25.6 million. And a Gartner survey in September 2025 found 62% of organisations had experienced a deepfake attack in the prior 12 months.
A note on scope: voice-deepfake and contact-centre figures from publishers such as Pindrop are real but measure a different channel, so we keep them separate from document-fraud metrics.
Identity and onboarding fraud
The sophistication shift is the story of 2025. Sumsub reported that identity fraud attempts actually fell slightly — to 2.2% of all verifications in 2025, from 2.6% in 2024 — yet the most sophisticated attacks rose 180% year on year. Sumsub defines sophisticated fraud as attacks combining multiple coordinated techniques — synthetic identities, deepfakes, layered social engineering, device tampering — which are far harder to detect than single-method schemes.
Professional services were a fast-rising target: Sumsub recorded a 232% year-on-year increase in identity fraud against legal, consulting, and accounting firms, attributed to their sensitive client data and reliance on manual onboarding. Among third-party fraud types, identity theft led at 28%, followed by account takeover at 18% and card testing at 17%.
Financial-document and email-based fraud
Document fraud and financial loss meet most directly in business payments. Per the FBI's Internet Crime Complaint Center (IC3) 2024 report, total reported losses reached a record $16.6 billion, up 33% on 2023, across roughly 860,000 complaints. Business Email Compromise (BEC) — which routinely turns on forged invoices, spoofed payment instructions, and fake authorisations — caused $2.77 billion in losses across 21,442 incidents, the second-costliest category; BEC losses totalled around $8.5 billion over 2022 to 2024. Cyber-enabled fraud accounted for roughly 83% of all reported losses, about $13.7 billion.
Looking forward, Deloitte's Center for Financial Services projects that generative-AI-enabled fraud in the US could rise from $12.3 billion in 2023 to $40 billion by 2027 — a projection, not an observed figure, but a widely cited indication of trajectory.
Regional snapshots
In the United States, record cybercrime losses of $16.6 billion in 2024 (FBI IC3), with BEC at $2.77 billion.
In Nigeria, financial institutions lost ₦52.26 billion to fraud in 2024, up from ₦17.67 billion in 2023 (NIBSS), with document- and identity-enabled fraud a major entry point. Our Nigeria document-fraud data hub covers this market in depth.
Across Africa, from analysis of more than 110 million identity checks, Smile ID found document fraud rose about 125% in the second half of 2024, and that by its 2026 report around 69% of biometric fintech fraud was AI-generated, with deepfake incidents up roughly sevenfold across 2024.
In the United Kingdom, Sumsub recorded a 94% year-on-year increase in deepfake attempts, consistent with the global sophistication shift.
Credential and qualification fraud
Document fraud extends well beyond finance. In Nigeria, the Federal Ministry of Education identified tens of thousands of people holding fake foreign degrees — about 21,600 from Benin and 1,105 from Togo between 2019 and 2023 — and suspended accreditation of degrees from both countries. Credential fraud, from fake degrees to forged professional certificates, is a global problem that has driven the growth of verification services for employers and universities.
The common thread: trust is moving to the source
Across every category, the same dynamic appears. Traditional document defences — a watermark, a signature, a familiar letterhead — were designed for a world where forgery was expensive. AI has ended that world. As the Entrust data shows, the centre of gravity has already shifted from physical to digital forgery, and detection alone is an arms race against tools that improve monthly.
The structural response is issuer-side verification: making a document confirmable at its source the moment it is created, so a recipient can verify authenticity in seconds rather than trying to detect an increasingly perfect fake. This is the model behind government registries, qualification-verification services, and document-authenticity platforms alike — and it is why so much of the response to AI-era fraud is converging on verify-at-source. VerifyDoc.ai applies that model to the documents organisations issue, attaching a QR-backed Certificate of Authenticity and a proof page so any recipient can confirm a document is genuine and unaltered. It does not replace KYC, fraud detection, or identity verification — it closes the gap that makes a document forgeable in the first place.
Methodology and sources
The figures on this page are drawn from publicly reported data as of mid-2026 and attributed to their originating publishers. Observed data and projections are labelled separately. Key sources include Entrust and Onfido's 2025 and 2026 Identity Fraud Reports, for digital-versus-physical forgery and deepfake frequency; Sumsub's Identity Fraud Report 2025 to 2026, for deepfake growth, the sophistication shift, and regional figures; the FBI's Internet Crime Complaint Center 2024 Annual Report, for US losses and BEC; the Deloitte Center for Financial Services GenAI fraud projection for 2023 to 2027; NIBSS's Nigeria industry fraud report for 2024; Smile ID's Digital Identity Fraud in Africa reports for 2025 and 2026; and the Federal Ministry of Education in Nigeria for the degree-milling figures.
Figures cover different scopes, methodologies, and periods, and are presented as reported. Always consult the original reports for full methodology.
Close the gap that makes documents forgeable
As AI collapses the cost of forgery, the documents your organisation issues are a target. VerifyDoc.ai lets you issue documents with a QR-backed Certificate of Authenticity that anyone can verify in seconds — turning your credibility into something you can prove. Start free or see how it works.
Related reading: Document fraud in Nigeria: the 2026 data and landscape and Detecting AI-generated bank statements.
This page compiles publicly reported third-party data for general information. Figures are attributed to their sources and periods; consult the original reports for full methodology.
