Who's really renting your car? Identity fraud, fake IDs and deepfakes in 2026

Fake driving licences, stolen identities and AI deepfakes increasingly target car rental. The 2026 threat picture, real cases and a layered defence — every figure sourced and attributed.

Think about what actually happens when you hand over a car. Someone shows a driving licence, maybe a passport, taps a card, and drives away in a vehicle managed by the rental company. The whole deal rests on one assumption: that the person is who the document says they are. A stolen or falsified identity can put that assumption to the test, whether the booking happens at a counter or remotely.

Identity fraud can involve a document altered by hand or a digitally fabricated image submitted remotely. It can mean a digitally fabricated document or a real stolen identity used with someone else's payment details. This guide sets out what's actually documented about identity fraud in car rental in 2026 — the numbers, the methods, the cases that reached a courtroom — and what genuinely works against it.

One thing up front, because it shapes everything that follows. A lot of the loudest numbers in this field come from companies that sell fraud-prevention software. That doesn't make them wrong, but it does mean you should know who's counting. So I label every source: law-enforcement or official, standards body, or commercial vendor. Where I only have a vendor's word for a figure, I say so. And where a number is genuinely solid — a guilty plea, a police operation — I lean on that instead.

In brief

  • One vendor saw an increase among its automotive customers. The share of fake IDs caught at automotive customers (car dealers and rental firms, combined) rose 21 % since December 2024, on more than 5 million IDs scanned in 2025 (vendor: IDScan.net, January 2026).
  • Digital fake IDs can be produced at scale. In February 2026, the U.S. Department of Justice reported a guilty plea involving OnlyFake, a service used to generate at least approximately 10,000 digital fake IDs. The case concerned identity checks at financial institutions; it does not measure fraud in car rental.
  • Law enforcement sees deepfakes going mainstream. Europol warns that deepfake technology could become a "staple tool" for organised crime, including document fraud by morphing faces to pass identity checks (official: Europol).
  • Onboarding checks are only one part of the process. In its mobility-industry analysis, verification vendor Sumsub reports that 70 % of fraud incidents occur after user verification. This is a vendor finding across mobility services, not a rental-only rate.
  • Assess checks in layers. Depending on the operation, document checks, face matching, presentation-attack detection and follow-up procedures can serve different purposes. Their effectiveness depends on the tools, test conditions and decisions made by the operator.

Why your counter is a target

Most fraud follows the money, and car rental sits on an unusually tempting pile of it. You routinely release a high-value, mobile, resaleable asset to someone you met minutes ago, on the strength of a document and a card. That combination — high value, high liquidity, low friction — is exactly what organised fraud looks for.

The playbook isn't mysterious. Verification vendors and law enforcement describe the same handful of patterns: rent a car with a stolen or fake identity and a stolen card, then drive it long distances, sell it abroad, or strip it for parts (vendor: Sumsub, "Trust on the Move", April 2026). Europol has documented criminal groups producing forged identity and travel documents for use across borders; the operation described in that source concerned migrant smuggling, not car rental. There's also a quieter, more everyday version: someone under the minimum age, or with a suspended licence, using an altered document to get behind the wheel.

None of this is unique to any one country, which is rather the point. The asset is mobile, the documents travel, and the fraud travels with them.

The threat in numbers — and who's actually counting

Here's the honest state of the evidence. Car-rental-specific fraud statistics are thin, and most of what exists comes from vendors measuring their own traffic. I'll give you the best figures available, with that caveat attached to each.

The most rental-adjacent hard number comes from IDScan.net, a document-verification vendor. Across its automotive customers — and this bundles car dealerships together with rental firms, so it is not rental on its own — the share of fake IDs detected rose 21 % since December 2024, drawn from more than 5 million IDs scanned in 2025 (vendor: IDScan.net, press release, 6 January 2026). It describes this vendor's customers, not a representative estimate for the entire sector or a rental-only rate.

On the documents themselves, Sumsub's mobility analysis breaks down the fakes it catches: of detected fraudulent documents, 72 % were ID cards, 13 % passports and 10 % driving licences (vendor: Sumsub, Identity Fraud Report 2025). The same source puts account takeover at 19 % of third-party fraud in 2025 across ride-hailing, car-sharing and micromobility, and — the figure I'd pin above your screen — reports that roughly 70 % of fraud incidents happen after the verification check in its mobility-industry analysis (vendor: Sumsub, April 2026). Sumsub does not provide a rental-only rate in that article. A completed check at pickup should not be treated as proof that every later risk has disappeared.

For context outside rental, an Experian report published in 2026 presents an online survey of 125 automotive dealers, conducted by HarrisX in September 2025. In that sample, 85 % reported suspected or confirmed auto-finance-related fraud at their organisation during the preceding 12 months: 30 % confirmed and 55 % suspected. These are dealer survey responses about finance fraud, not a measured rate for rental agencies.

You'll also see eye-watering deepfake projections in the trade press — single-vendor forecasts of several-hundred-percent jumps in a single year. I'm leaving those out. They're unverifiable, they come from one company each, and they don't survive the editorial test for this blog. The qualitative warning from Europol below does the same job without the false precision.

How the fraud really works

Strip away the jargon and there are four recurring methods. They're not exotic; they're industrialised.

The altered or counterfeit physical document. The classic: a tampered or wholly fake driving licence presented at the desk. What's changed is the supply. The OnlyFake service produced digital images of identity documents, including driving licences and passports. Its operator pleaded guilty in the United States; according to the U.S. Department of Justice, the service was used to generate at least approximately 10,000 fake IDs between roughly 2021 and 2024. The official case concerns attempts to bypass checks at financial institutions; it is an example of document supply, not a statistic about rental fraud.

The stolen-but-real identity. Here the document is genuine, or genuinely good, and belongs to a real person who isn't in the room ; it can also be used alongside stolen payment details. A genuine document alone does not establish that the person presenting it is its rightful holder. The prevalence of this specific method in rental losses is not established by the sources cited here.

The synthetic identity. A fabricated person, stitched together from a mix of real and invented details, built up over time to look legitimate. Sumsub reports that synthetic-identity attack rates in auto lending in 2025 were almost double their 2020 level (vendor: Sumsub) — adjacent to rental, and a sign of where patient, organised fraud is heading.

The deepfake or morphed face. As more verification moves to a phone camera, the attack moves with it. Europol's innovation lab flagged that deepfakes and digitally manipulated facial images open a new route for document fraud, combining or "morphing" two faces so a single photo can pass for two different people, and warned the technology could become a staple tool for organised crime (official: Europol, Facing reality?, 2022). Its EU Serious and Organised Crime Threat Assessment goes further, describing criminal networks using AI to automate fraud at scale (official: Europol, EU-SOCTA 2025). The cutting edge isn't holding a photo up to a camera anymore; it's injecting a synthetic video straight into the verification feed.

What it looks like when it reaches a courtroom

Abstract numbers land better with a concrete case or two. These are vehicle-fraud prosecutions rather than rental specifically, but they show the same engine at work — a fake driving licence used to get a car that isn't yours.

In November 2025, Hits Radio reported on a Surrey Police investigation into a group that targeted 50 vehicles valued at more than £1 million between February 2022 and April 2024. The report describes vehicles being reidentified and sold to dealerships using false documents. In a separate case outside car rental, Hertfordshire Constabulary reported in April 2024 that 16 people were sentenced in a car-finance fraud involving 75 vehicles valued at more than £2 million. The police said some collectors were given fake identity documents. These examples concern vehicle sales and finance, not rental-specific fraud.

On the document-supply side, the European Commission reported in October 2026 that French and Spanish authorities, supported by Europol, had discovered a suspected forged-document production facility in Spain. That investigation concerned suspected migrant smuggling, not rental fraud. It illustrates a documented source of forged documents without establishing that any particular rental incident originated there.

Stopping it: defence in layers

No single check stops a determined, well-equipped fraudster. The thing that works is layering, so that a method which beats one control runs into the next. Here's the stack, from document to behaviour.

Authenticate the document, not just glance at it. Modern verification reads a document's security features, checks the machine-readable zone, and — on e-passports and the newer e-ID cards — reads the NFC chip, which is far harder to forge than the printed surface. AI-assisted forgery detection looks for the tell-tale artefacts of generated or edited images. This is the layer the cheap fake-ID services are built to beat on sight, which is exactly why it can't be the only one.

Compare the document with the person presenting it. Face matching can compare a document portrait with an image of the customer. Presentation-attack detection, often called liveness detection, is designed to identify some attempts to fool a biometric check, but a result is not a guarantee that a person is genuine or that every type of manipulation has been caught. In a 2023 NIST evaluation, none of the tested algorithms detected every type of presentation attack in the study.

Testing can make those limits easier to understand. ISO/IEC 30107-3 concerns testing and reporting for presentation-attack detection; iBeta publishes reports for specific products and versions. Ask a provider for the test report, the version tested, the attack types and the error rates, and ask how other threats such as injected media are handled. A laboratory result is evidence about the stated test, not proof that every real-world attack will be detected.

Check the identity against the wider world. Depending on your market and obligations, that can mean validating the driving-licence number and entitlement where an official check exists, screening against sanctions or known-fraud lists, and flagging duplicate or previously-abused identities. Device, network and behavioural signals — a freshly created account, a mismatched location, an impossibly fast form-fill — add context the document can't.

Don't stop at pickup. In its mobility-industry analysis, Sumsub reports that roughly 70 % of fraud incidents happen after verification; that vendor figure should not be read as a rental-only rate. A held deposit or card authorisation, telematics and geofencing that notice a car heading for a port it has no business visiting, and a real return process all extend the defence across the whole rental, not just its first thirty seconds.

Back it with people and habits. The counter still matters. Staff who are trained to feel a document, to check the name on the card against the name on the licence, and to register the small wrongness of someone in a hurry to skip a step — they catch things software misses. Decide what information is necessary for your process, who can access it and how long it should be kept. Applicable rules depend on the jurisdiction and the processing; France's CNIL explains the principles of data minimisation and limited retention.

Where PASS2RENT fits, and where it doesn't

Several of those layers are exactly what a contactless rental platform is built to run. PASS2RENT's driver app handles identity verification with document checks, a biometric face-match and liveness detection before a car is released, captures an eight-point photo inspection that ties the renter and the vehicle's condition to the contract, and takes an e-signature; the admin console keeps driver verification, contracts and fleet telematics in one place, so the "after pickup" layer isn't an afterthought.

Now the honest part, because I'd rather you trust the rest of the article. No verification system catches everything, and anyone who promises a 100 % fraud-proof check is selling you the very overconfidence fraudsters rely on. A platform can help operators apply the checks they choose; it does not abolish risk. Nor is PASS2RENT a legal, insurance or compliance adviser — your fraud policy, your KYC obligations and your acceptance decisions remain yours. Depending on its configuration and the checks available, software can help an operator organise parts of that process, including for bookings scheduled outside desk hours.

The bottom line

A rental agency cannot remove every risk, but it can make its checks and escalation process clear. Independent tests help operators understand the capabilities and limitations of the tools they choose. You need document authentication, a face-match and presentation-attack checks whose limitations are understood, appropriate additional checks, and a procedure for following the rental after the vehicle is handed over. The right combination will depend on the agency, its customers and the rules that apply to its operations.

Want to explore the customer journey? See the PASS2RENT driver app.


FAQ

How common is identity fraud in car rental, really?

There is no clean, rental-only public figure — most data comes from verification vendors measuring their own traffic, which is why I flag the tier on each number. The most rental-adjacent signal is from IDScan.net: fake IDs detected at automotive customers (dealers and rental firms together) rose 21 % since December 2024, on 5 million-plus IDs scanned in 2025 (vendor: IDScan.net, January 2026). Treat it as a trend for the sector, not a precise rental rate.

Are AI-generated fake IDs actually a real threat or just hype?

Digital fake documents are a documented concern. In February 2026, the U.S. Department of Justice reported a guilty plea involving OnlyFake, which generated at least approximately 10,000 digital fake IDs. The case concerned financial identity checks, not car-rental incidents; it does not establish how common the problem is among rental customers.

What is "liveness detection" and why does it matter?

It describes checks designed to detect certain attempts to present a false biometric sample, such as a photograph or a mask. Results depend on the system and the attack. NIST found that none of the algorithms in its cited evaluation detected every presentation attack tested. Ask providers for the test scope and results, and how errors or uncertain cases are handled.

Is checking the ID at pickup enough?

An initial check does not eliminate every later risk. Sumsub reports that roughly 70 % of fraud incidents occur after verification in its mobility-industry analysis; this is not a rental-only statistic. A deposit or card authorisation, telematics and geofencing during the rental, and a proper return process extend the defence across the whole booking.

We're a small agency — isn't this only a big-operator problem?

The asset and the method don't care about your size. A car is a car, and a fake driving licence works the same at a two-branch agency as at an airport desk. The good news is that the layered defence — document checks, independently evaluated presentation-attack detection, appropriate additional checks and procedures during the rental — is now available as software rather than a fraud team, which is what puts it within reach of a small operator.

The answers in this FAQ rely on the same sources as the rest of the article (see below).


Related

Sources

Official / law-enforcement

Independent reporting

Standards bodies and data protection

Commercial vendors (figures reported by companies that sell verification/fraud tools — read as such)

Vendor figures retain their stated dates. The NIST and data-protection sources added in November 2026 provide context on the limitations of biometric checks. This article is for information and professional education; it is not legal, insurance or compliance advice. Figures from commercial vendors are attributed to their source and date and do not carry the weight of an official statistic; vendor projections that could not be independently verified have been deliberately left out. Verify the latest data before making decisions.

P2R

PASS2RENT Team

Our team of experts shares insights, tips, and best practices to help you succeed in the car rental industry. Stay tuned for more valuable content!

Topics:Car RentalBusiness
Share:

Want to read more?

Explore more articles from this category and stay updated with the latest insights.

View More Articles