Aviva detects record £233m in AI-generated insurance fraud claims in 2025
Aviva detected a record £233 million in fraudulent insurance claims in 2025, with fraudsters increasingly using AI-generated images and manipulated documents to support fake motor, liability and home insurance claims. The insurer uses AI tools and advanced analytics to identify and stop suspicious claims, and 37 years of custodial sentences were secured for serious fraud offences.
What happened
Aviva identified more than 18,400 suspect insurance claims across its brands in 2025, with a combined value of £233 million. The fraud claims level was a record for the insurer, although this was the first year that it included the Direct Line brands it acquired.
Fraudsters used AI-generated images and manipulated documents to support claims, particularly in motor insurance, fabricating accident scenes and damage imagery. Motor insurance fraud accounted for more than seven in 10 cases detected, with the value of detected motor fraud rising 39% year-on-year.
Reported impact
Pete Ward, head of claims counter fraud at Aviva, stated that fraud "drives up the cost of insurance for everyone." The insurer detected more than £638,000 of fraud stopped every day in 2025.
Index analysis
Fraudsters moved away from staged collisions towards exaggerated claims for vehicle damage, repair costs, credit hire and injury, often using wider cost pressures as justification. A similar pattern emerged in liability insurance, where the value of fraudulent claims detected increased 32%, and home insurance fraud rose 15%.
Governance controls
In response to the growing sophistication of fraud, Aviva used AI tools and advanced analytics, overseen by humans, to detect suspicious claims faster. In 2025, 37 years of custodial and suspended sentences were secured for the most serious fraud offences across the Aviva and Direct Line brands.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.