Statistics · October 2026

Visual fraud statistics. The 2026 figures, with their sources.

We gathered on a single page the most recent public figures on fraud with edited, recycled or AI-generated photos. Every figure has its source, so you can verify and cite it.

Vouch teamOctober 23, 20267-minute read

In short

  • 36% of U.S. consumers would consider altering an image to strengthen a claim.⁠1
  • With images from the most advanced generators, people do no better than flipping a coin.⁠2
  • Refund abuse is the most widespread fraud in online commerce: 41% of merchants suffered it.⁠3
  • UK insurers detected £1.34 billion in fraudulent claims in 2025.⁠4

Photo fraud: the big picture

36%

of U.S. consumers would consider altering an image or document to strengthen a claim. Among Gen Z, 55%.⁠1

Survey of 1,000 consumers, Dec 2025–Jan 2026.
300%

increase between 2021-22 and 2022-23 in cases with images, videos and documents altered with apps detected by Allianz UK.⁠5

Allianz UK, 2024.
80%+

of fraud attacks on merchants already use AI, and claims with fabricated damage are the fastest-growing form of return abuse.⁠6

Forter report on UK merchants, Jun 2026.
$1.87B

is the value of the fake image detection market in 2026. It could reach $7.43 billion by 2031.⁠7

Mordor Intelligence estimate, with annual growth near 32%.

How well people detect it

48.5%

accuracy for 136 people judging evidence photos made with the most advanced AI generators. Across all images, 64.8%.⁠2

Preprint from Université de Montréal and McGill, Jun 2026.
49.4%

accuracy for study participants classifying images as real or AI-generated.⁠8

Communications of the ACM, 2025.
62%

accuracy across 287,000 judgments by more than 12,500 people in a real-vs-fake image game.⁠9

Microsoft Research, 2025.
32%

of claims professionals feel very confident they can identify a deepfake. 98% believe AI editing is driving fraud.⁠1

Survey of 300 claims professionals, Dec 2025–Jan 2026.

More detail in the guide to telling whether a photo was generated or edited with AI.

Insurance

£1.34B

in fraudulent claims detected by UK insurers in 2025, across 93,900 cases. More than half were motor.⁠4

Association of British Insurers data.
£14,300

was the average fraudulent claim in the UK in 2025, the second highest on record.⁠10

Association of British Insurers data.
R$ 1.1B

in proven fraud in Brazil in 2024, out of R$ 5.4 billion in suspicious claims.⁠11

CNseg Fraud Quantification System.
3.3%

of earned premiums in Brazil was proven fraud in 2025, up from 1.8% in 2021.⁠12

Revista Apólice, Jul 2026.
60%

of the fraud cases AXA detected in Spain in 2024 were real claims with added damage.⁠13

AXA Spain, 2025.
+71%

growth in fraud detected by Admiral in 2025, to £86.8 million. Cases included car photos edited with a different number plate.⁠14

Admiral Group, 2026.
£233M

across 18,400 suspicious claims uncovered by Aviva in 2025.⁠15

Aviva, Jun 2026.
$308.6B

is the yearly cost of insurance fraud in the United States.⁠16

Coalition Against Insurance Fraud, 2022.

More detail in the insurance fraud report.

Delivery and refunds

41%

of online merchants suffered refund and policy abuse in the past year: the most widespread type of fraud.⁠3

Survey of 1,278 eCommerce professionals in 37 countries, Nov–Dec 2025.
61%

of merchants saw refund abuse grow in the past year. 20% reported increases of more than 25%.⁠3

Same survey.
52%

of merchants suffered false “item not received” claims, the most common form of refund abuse.⁠3

Same survey.
48%

of consumer-side fraud on delivery and ride-hailing platforms was refund abuse.⁠17

Cases detected in 2024.
73%

of gig economy fraud teams say stopping multi-accounting is their priority.⁠18

Incognia, 2025.

More detail in the delivery fraud report.

E-commerce and returns

44%

of merchants received returns of used or damaged products. 29% received counterfeit products.⁠3

Survey of 1,278 eCommerce professionals in 37 countries, Nov–Dec 2025.
$849.9B

in merchandise U.S. consumers will return in 2025: 15.8% of sales. Online, 19.3%.⁠19

National Retail Federation estimate.
$100B

in preventable losses from return fraud and abuse in retail in 2025, 14.2% of all returns.⁠20

Appriss Retail, Feb 2026.
35,000

subscribers on the Telegram channel of REKK, a ring that sold fraudulent refunds for a 30% commission.⁠21

Amazon lawsuit, Dec 2023.
12.2%

of their fraud chargebacks and disputes are won by Latin American merchants, versus 16.7% globally.⁠3

Merchant Risk Council, 2026.

More detail in the e-commerce return fraud report.

Warranties

Up to 15%

of an average company’s warranty costs can come from fraudulent claims. Those costs usually equal 1% to 4% of sales.⁠22

Warranty Fraud Management, Wiley, 2016.
10%

of more than 75,000 auto warranty verifications was flagged as highly suspicious: $8.1 million in potential fraud.⁠23

Truepic data, an image verification vendor.
100,000

warranty claims verified by OtterBox with authenticated photos in under four months.⁠24

Truepic data.

More detail in the warranty fraud report.

Technology and regulation

10B+

pieces of content already carry Google’s invisible watermark. It only works for content made with its tools.⁠25

Google, May 2025.
Half

of fake images were detected by Taobao and Tmall’s after-sales model, according to reports.⁠26

OECD.AI, Jan 2026.
Aug 2026

the EU AI Act’s transparency obligations on AI-generated content took effect.⁠27

European Union.
2021

the C2PA content credentials standard was created, backed by Adobe, Arm, BBC, Intel, Microsoft and Truepic.⁠28

C2PA.

What they are and what they’re for, in the article on C2PA.

How Vouch solves it

Vouch is a visual fraud detection platform. It analyzes every photo in under 30 seconds with three layers: detection of images generated or edited with AI, a search for the image online and in previous claims, and the customer’s context. It returns a 0–100 risk score with the reason, and gray-area cases go to a specialist. It works from Vouch Web, with no integration, or via API.

Frequently asked questions

How many people would alter a photo to win a claim?

In the United States, 36% of consumers would consider altering an image or document to strengthen a claim, and among Gen Z the figure reaches 55%, according to a 2026 Verisk survey.

Can people tell a real photo from an AI-made one?

Less and less. In a study published in Communications of the ACM, accuracy was 49.4%, and in a 2026 study with evidence photos, accuracy dropped to 48.5% with images from the most advanced generators: the same as flipping a coin.

What is the most common fraud in online commerce?

Refund and policy abuse: 41% of merchants suffered it in the past year, according to the Merchant Risk Council. Within that abuse, the most common form is false “item not received” claims, which affected 52% of merchants.

How much does insurance fraud cost?

In the UK, insurers detected £1.34 billion in fraudulent claims in 2025. In Brazil, proven fraud reached R$ 1.1 billion in 2024. In the United States, the Coalition Against Insurance Fraud estimates a cost of $308.6 billion a year.

Can I cite these statistics?

Yes. Cite the original source of each figure, linked at the end of the page, and this page as a compilation if you wish.

Sources

This page is based on public sources only; every figure can be verified.

  1. Verisk. AI Editing Tools Are Fueling a New Era of Insurance Fraud, According to New Research from Verisk (Mar 2026)
  2. arXiv (Universidad de Montreal y McGill). Can You Trust What You See? Human and AI Detection of Synthetic Legal Evidence (Jun 2026)
  3. Merchant Risk Council. 2026 Global eCommerce Payments & Fraud Report (Mar 2026)
  4. Bodyshop Magazine (datos de la ABI). Fraudulent insurance claims worth £1.34bn detected in 2025 (2026)
  5. PropertyCasualty360. Fraudsters using AI to manipulate images for false claims (May 2024)
  6. Retail Rewired. Forter: Retailers report rising returns fraud, with 80% of fake claims fuelled by AI evidence (Jun 2026)
  7. Mordor Intelligence. Fake Image Detection Market Size and Share Research Report, 2031 (2026)
  8. Communications of the ACM. As Good as a Coin Toss: Human Detection of AI-Generated Content (Oct 2025)
  9. Microsoft Research. How good are humans at detecting AI-generated images? Learnings from an experiment (2025)
  10. Insurance Business. Fewer fraud cases, fatter claims as average scam hits £14,300 (2026)
  11. Sindicato das Seguradoras. Fraudes em seguros somam R$ 1,1bi: como se proteger? (Aug 2025)
  12. Revista Apólice. Fraudes no seguro atingem maior nível da série histórica (Jul 2026)
  13. Bolsamanía (Europa Press). Axa evita 87 millones en fraude al seguro en 2024, un 3,6% más (2025)
  14. Admiral Group. The White Lie Effect: 1 in 8 people admit to exaggerating insurance claims (2026)
  15. Aviva plc. Aviva stops record levels of claims fraud as scams become more sophisticated (Jun 2026)
  16. Coalition Against Insurance Fraud. The Impact of Insurance Fraud on the U.S. Economy (2022)
  17. Incognia. Frontline Report: Gig Economy Edition (Feb 2025)
  18. Incognia. The State of Fraud in the Gig Economy 2025 (Oct 2025)
  19. National Retail Federation. Consumers Expected to Return Nearly $850 Billion in Merchandise in 2025 (Oct 2025)
  20. Appriss Retail. 2026 Total Retail Loss Benchmark Report (Feb 2026)
  21. Sourcing Journal (WWD). Amazon Lawsuit Names 27 Defendants Running Global Refund Fraud Scheme (Dec 2023)
  22. Wiley (Kurvinen, Töyrylä y Murthy). Warranty Fraud Management: Reducing Fraud and Other Excess Costs in Warranty and Service Operations (2016)
  23. F&I and Showroom (datos de Truepic). Why the auto warranty industry needs image data authentication
  24. Truepic. Truepic and OtterBox partner on warranty claims
  25. Google. SynthID Detector — a new portal to help identify AI-generated content (May 2025)
  26. OECD.AI (AI Incidents Monitor). Taobao and Tmall Deploy AI to Combat Fraudulent Refunds with Fake Images (Jan 2026)
  27. Jones Walker LLP. The EU approved a high-risk AI delay, but most transparency obligations remain (2026)
  28. C2PA. Technology and media entities join forces to create standards group aimed at building trust in online content (Feb 2021)