Zurich UK detects a record £109.3m of fraudulent claims as gangs turn to generative AI

One bogus claim every two hours. That is the pace of fraud Zurich UK detected over the last 12 months: 4,359 fraudulent claims worth £109.3 million.^[1][2]

Zurich also says fraudsters are making growing use of generative AI to create claim documents and evidence.^[1]

What Happened

Zurich UK published the figures on 5 November 2026. The 4,359 claims are its highest ever by value and volume.^[2] Detected fraud claims are up 22% from 3,562 in 2022.^[1]

The split by line of business:^[3]

• Casualty: £61.6 million across 2,380 claims
• Motor: £28.1 million across 1,027 claims
• Property: £19.6 million across 951 claims

Zurich lists generative AI alongside organised networks, ghost broking, crash-for-cash scams and fictitious accidents as signs that fraudsters are becoming increasingly sophisticated.^[1] Its release gives no count of how many claims involved AI-generated material.

"Insurance fraud in the UK is becoming more sophisticated, more organised and increasingly technology-enabled," said Scott Clayton, Head of Claims Fraud at Zurich UK and Chair of the Insurance Fraud Bureau (IFB).^[2] He named staged accidents, ghost broking and AI-generated evidence among the tactics criminal networks use.^[1][2]

Connected AI and advanced analytics identified 860 fraudulent claims worth £17.9 million, around one in five fraud cases detected by volume.^[1]

Zurich uses Quantexa technology to link claims with Companies House, Equifax, industry intelligence and the Insurance Fraud Register.^[2] Zurich said it was the first insurer to connect every claim across multiple data sources for an individual claimant.^[2]

Zurich described several cases:

• A routine workplace accident claim that connected analytics tied to significant debts and suspicious whiplash claims.
• Several seemingly unrelated claims tied to the same accident location.
• A claimant using an alternative identity to hide a claims history.^[1]
• People living in one part of the UK linked to suspected staged accidents hundreds of miles away.^[2]

"Organised fraud doesn't stop at one insurer's front door, so defences can't either," said Vishal Marria, Founder and CEO of Quantexa. "Left unchecked, the cost lands on honest customers through higher premiums and slower service."^[1]

Why It Matters

A claims file is only as good as the evidence in it. If documents and evidence can be generated, a damage photo or a supporting record may show something that never happened.

The image looks real. The evidence might not be.

Zurich is one insurer. Across the industry, the Association of British Insurers says insurers detected £1.34 billion of fraudulent claims in 2025. The total value rose 14% even as case numbers fell slightly. Insurers also stopped £1.8 billion of fraudulent applications, with application fraud cases up 24%.^[2]

Zurich pairs data with people. "The combination of advanced analytics, specialist investigators and human judgement is proving a powerful defence," Clayton said.^[1][3]

What to Do

• Treat claim photos and documents as evidence to check, not as proof on arrival.
• Ask for original files, not screenshots. Compare them with other evidence.
• Look for repeats: the same image, location, solicitor or repairer across unrelated claims.
• Send anything flagged to a specialist investigator. A flag is a reason to look closer, not a finding of fraud.
• In the UK, report suspected insurance fraud to the Insurance Fraud Bureau's CheatLine.^[3]

Verify What You See and Hear

Upload any suspicious image or video frame at witness.vision/scan to check for signs of AI generation. The Witness Chrome extension scans images directly on any webpage.

Witness gives a detection signal, not proof. An image flagged as potentially AI-generated needs human review. A clean result does not prove a photo is authentic.

Scan now at witness.vision/scan