Nearly all insurers - 98% - report that AI editing tools are fueling digital fraud, yet only 32% say they are very confident in detecting these deepfakes, according to the Verisk State of Insurance Fraud Study. The result: claim photos and documents can no longer be trusted at face value, forcing the industry to rethink how it evaluates evidence.
Fraudsters are using AI to create or alter digital images, invoices, and receipts that look authentic. Traditional claims managers, who are rarely trained in digital forensics, must now assess not only the facts of a claim but also whether the supporting documentation has been manipulated.
Why manual fraud review falls short
AI-generated images have moved beyond obvious tells like six-fingered hands or "Uncanny Valley" distortions. Even experienced adjusters can be fooled. Aviva's latest fraud data illustrates the scale: the insurer detected about $311 million in suspected claims fraud in 2025. Aviva warned that fraudsters increasingly use AI-generated images and manipulated documents to support false or exaggerated claims.
As a result, "claims teams are not just reviewing the facts of the claim but also evaluating whether the supporting evidence has been altered," the insurer said. That added layer of complexity means every piece of digital evidence now demands scrutiny.
When AI creates more noise and risk
Many insurers are adopting AI fraud scoring tools to flag suspicious claims. In practice, however, these tools can flood teams with alerts or black-box scores that offer no clear evidence of what triggered the flag. Adjusters are left to determine whether the score is accurate - while still facing the same volume of claims.
That creates legal exposure. If a claimant challenges a denial, the insurer must provide a defensible paper trail. "If their only basis for denying a claim is a black box score, then the insurer will be under major scrutiny and exposure," the Verisk study noted. A score alone is not enough when AI is being used to manufacture evidence.
Explainable AI becomes critical. Carriers need tools that surface clear evidence - connecting data points across images, timelines, claimant behavior, and records - so adjusters can act with confidence. That is where AI for Insurance training can help teams understand how to evaluate and trust these outputs.
From faster flags to defensible decisions
"AI does well at gathering data, summarizing information, identifying patterns, and routing claims to the right person," the study states. But insurers need tools that provide a full picture, not just a single flag. This means connecting findings to the evidence behind them so adjusters have a clear trail if the decision is challenged.
One photo can be faked; it is much harder to fake the broader story across multiple documents, timelines, and claimant behavior. AI's real value is in helping claims teams spot inconsistencies across that broader story before questionable claims move further through the process.
Why this matters for insurance professionals
Fraud detection tools that produce black-box scores add operational burden and legal risk. Insurance professionals should demand fraud scoring systems that provide clear, explainable evidence - not just a number. Without that, adjusters cannot confidently accept or deny a claim, and the company remains exposed to costly litigation and reputational damage. The goal is AI that supports human judgment, not replaces it, and leaves a defensible paper trail behind every decision.
Your membership also unlocks:
