About this report: published by our Trust and Safety Research Team as part of our ongoing coverage of identity fraud and document verification trends. If identity fraud is affecting you personally, contact the FTC at IdentityTheft.gov or your local law enforcement agency.
"Fake ID" covers two almost unrelated stories in 2026, and most coverage blends them into one.
There is the story that has not changed much in fifteen years: a college student, a bar, a birthday still a few months away. And there is the story that barely existed three years ago, which is AI-generated documents realistic enough to pass automated checks, now serious enough for its own section in the FBI's cybercrime report and binding new law across three continents.
We grouped the year's most credible data into five categories (federal enforcement, AI and synthetic fraud, campus research, physical smuggling, and global regulation) so each claim sits next to the source that backs it.
Get a Scannable Fake ID That Passes Every Check
Federal Enforcement
The clearest sign that fake IDs became a federal cyber security story in 2025 is where the data now lives.
The FBI's 2025 Internet Crime Report, compiled by the Internet Crime Complaint Center (IC3), added a dedicated AI-fraud category for the first time in the report's roughly 25-year history.
IC3 recorded 22,364 AI-related fraud complaints, with losses near $893 million. Government impersonation complaints, a category IC3 links directly to fraudulent identification documents, nearly doubled year over year to $797.9 million. The report names fake ID documents specifically as one of the AI-generated assets behind the category.
The year's benchmark prosecution came in February 2026, when the U.S. Department of Justice charged the creator of OnlyFake, an underground platform that DOJ charging documents say generated more than 10,000 fraudulent digital documents across all 50 U.S. states and 56 countries, including driver's licenses, passports, and Social Security cards sold for cryptocurrency.
Independent journalists had already shown a convincing synthetic ID could be produced through the platform in minutes, a demonstration that reshaped how identity-verification vendors evaluate document checks industry-wide.
AI and Synthetic Identity Fraud
This is where the growth is, even though the current share is still small.
Sumsub's Identity Fraud Report 2025-2026 found AI-generated documents made up 2 percent of all fake documents detected globally, measured over roughly six months since this document type was first detected in April 2025.
Sumsub frames the number as an early rate, not a ceiling. Sophisticated fraud overall (a broader category including but not limited to AI documents) rose from about 10 percent of fraud attempts in 2024 to 28 percent in 2025, a 180 percent increase.
Shufti's Identity Fraud Index projects deepfake identity fraud will grow 495 percent in 2026 versus 2025, with document-based deepfakes specifically projected to grow nearly 40-fold. These are forward projections from a named industry tracker, not confirmed historical totals.
Entrust's 2026 Identity Fraud Report, built from more than 1 billion identity verification events across 195 countries, found deepfakes now account for one in five biometric fraud attempts globally.
Deepfake selfie attempts rose 58 percent year over year, and injection attacks (feeding manipulated images directly into a verification system rather than presenting them to a live camera) rose 40 percent.
Campus Research
Set against the AI-fraud numbers above, the academic research on college fake ID ownership looks almost stationary.
A national freshman survey, cross-sectional across multiple U.S. universities and published in a peer-reviewed addiction research journal, found 7.7 percent of incoming students already owned a fake ID before starting college.
A separate study of 1,454 underage students at a large Southeastern university found 38.5 percent had owned one at some point during college, peaking at 39 percent during the third year. A third study, of 1,098 underage students at a large Midwestern university, found 21.0 percent currently owned a fake ID. Of those, 93.5 percent had used it, and only 29.1 percent reported ever being caught.
That last figure is worth sitting with on its own. Roughly seven in ten self-reported fake ID uses, in this study, went undetected entirely.
That is self-report data, not law enforcement records, so it measures perceived rather than confirmed detection risk. But it goes a long way toward explaining why campus-level ownership has held steady for over a decade despite consistent enforcement: the felt risk stays low, even when the legal risk does not.
None of these three studies show a sharp recent increase. The real year-over-year movement is happening in the AI-fraud category above, a category that sits entirely outside how campus surveys are designed to measure this issue.
Physical Smuggling
Not everything moved online. U.S. Customs and Border Protection continues to report a seasonal pattern its own press releases call the "second semester surge," with physical counterfeit ID seizures spiking every year around the start of the academic calendar.
In one recent reporting period, CBP officers at Chicago's O'Hare International Airport confiscated 984 counterfeit driver's licenses in a single week in January.
CBP's own seizure data, corroborated by industry reporting, has placed the share of physical counterfeits manufactured in China or Hong Kong as high as 97 percent in past reporting years, with shipments concealed inside laptop chargers, plastic mirrors, and jewelry boxes.
Global Regulation
Fraud is not distributed evenly worldwide, and neither is the regulatory response to it.
Sumsub's regional breakdown found fraud rates fell 14.6 percent in North America and 5.5 percent in Europe during 2025, a trend the World Economic Forum, citing the same dataset, attributes to more mature regulatory ecosystems and stronger verification adoption in both regions.
Elsewhere moved the opposite direction: Africa up 9.3 percent, Latin America and the Caribbean up 13.3 percent, Asia-Pacific up 16.4 percent, and the Middle East up 19.8 percent, the steepest increase recorded anywhere.
Veriff's Identity Fraud Report 2026 complicates the European picture. It found the EU and UK's annual mean fraud rate rose nearly 2.3 times year over year, in apparent tension with Sumsub's European decline.
Veriff's own report flags a likely explanation: the increase may partly reflect newly measured activity following wider adoption of anti-fraud detection tools, meaning more fraud is being caught and logged rather than more fraud actually occurring. It is a useful reminder generally, because fraud statistics measure detected fraud, and detection capacity is itself a moving variable.
Regulators, meanwhile, moved from proposal to binding enforcement across several major markets during this period:
- EU Artificial Intelligence Act. Binding obligations now apply to developers and operators of high-risk AI systems, including identity verification tools.
- EU Digital Identity Wallet (eIDAS 2.0). Scheduled to reach every EU citizen, resident, and business by the end of 2026.
- NIST Digital Identity Guidelines (SP 800-63-4). The US federal government's updated verification standard, with revised authentication and anti-spoofing rules.
- INTERPOL's Project SynthWave. Actively training law enforcement across Southeast Asia to recognize deepfake evidence in criminal cases.
- UK Economic Crime and Corporate Transparency Act 2023. Made "failure to prevent fraud" a standalone corporate liability offense.
- EU Digital Services Act inquiries. In September 2025, EU authorities opened formal inquiries into Google, Microsoft, Apple, and Booking.com over platform fraud prevention, per Veriff's report.
Key Takeaways
| Finding | Source and figure |
|---|---|
| AI fraud became its own federal category | FBI 2025 Internet Crime Report: 22,364 complaints, $893 million in losses |
| A single platform supplied documents at scale | February 2026 OnlyFake prosecution (DOJ): 10,000+ documents, 50 states, 56 countries |
| AI documents are still a small share, but growing | Sumsub: 2 percent of detected fakes over six months; large projected growth per Shufti and Entrust |
| Campus ownership has been broadly stable | Peer-reviewed studies: 7.7 percent to 38.5 percent depending on population and timing |
| Most self-reported use went undetected | Midwestern university study: only 29.1 percent said they had ever been caught |
| Physical seizures follow a seasonal pattern | CBP: 984 counterfeit licenses at O'Hare in a single January week |
| Regional fraud trends split sharply | Sumsub: North America down 14.6 percent, Europe down 5.5 percent, Asia-Pacific up 16.4 percent, Middle East up 19.8 percent |
| Two reports disagree on Europe | Veriff found EU/UK fraud up 2.3 times, partly attributed to improved detection rather than more fraud |
| Regulation moved from proposal to enforcement | EU AI Act, EU Digital Identity Wallet, NIST SP 800-63-4, UK corporate liability law |
This report will be updated as new data becomes available. Figures reflect the most recent published data as of August 2026.
Ready to Order Your Fake ID?
Frequently Asked Questions
What changed in the FBI's 2025 Internet Crime Report?
IC3 added a dedicated AI-fraud category for the first time in roughly 25 years of publishing the report. That category logged 22,364 complaints and losses near $893 million, and the report names fraudulent identification documents as one of the AI-generated assets behind it.
How common are AI-generated fake documents right now?
Still a small share of the total. Sumsub measured them at 2 percent of all fake documents detected globally over a six-month window starting April 2025. The number worth watching is the growth rate rather than the current share, since sophisticated fraud overall jumped from about 10 percent to 28 percent of attempts in a single year.
Has college fake ID ownership actually increased?
Not by any measure in the peer-reviewed data. Ownership sits somewhere between 7.7 percent of incoming freshmen and 38.5 percent of students at some point during college, depending on the study, and those figures have held broadly steady for over a decade. The real movement in 2026 is in the AI-fraud category, which campus surveys were never designed to capture.
Why do two major reports disagree about fraud in Europe?
Because they may be measuring different things. Sumsub recorded a 5.5 percent decline in Europe while Veriff recorded a 2.3 times increase across the EU and UK. Veriff's own authors suggest the rise partly reflects wider adoption of detection tools, meaning more of the existing fraud is being caught and logged rather than more fraud happening.
Are physical counterfeit IDs still being seized in volume?
Yes, and on a predictable schedule. CBP describes a "second semester surge" every year around the start of the academic calendar, and officers at O'Hare alone confiscated 984 counterfeit driver's licenses in one January week. Past reporting years have put the share manufactured in China or Hong Kong as high as 97 percent.
What regulation should someone actually pay attention to?
The EU Artificial Intelligence Act and the EU Digital Identity Wallet carry the widest reach, since both apply directly to identity verification systems and are on a 2026 timeline. In the US, the updated NIST Digital Identity Guidelines set the federal verification standard, including revised anti-spoofing rules.
Final Thoughts
The most useful thing in this report is not any single figure. It is the gap between the two stories. Campus ownership has barely moved in fifteen years, while the automated, synthetic side of document fraud went from a niche concern to its own federal reporting category in about three.
Those two trends get written about as if they were one topic. They are not, and treating them as one is how coverage ends up either overstating a stable problem or understating a fast-moving one.
It is also worth remembering what every number here actually measures. Fraud statistics count detected fraud, and detection capacity keeps improving. A rising figure can mean more fraud, better catching, or both, which is exactly the ambiguity Veriff flagged in its own data.
Researched and written by our Trust and Safety Research Team for educational and public-awareness purposes. We do not encourage the use of identity documents for any type of illegal or fraudulent activity.