Insurance fraud is undergoing a structural shift as generative AI tools rewire how false claims are filed, with carriers under pressure to adapt or absorb mounting losses, according to fresh analysis from Gen Re developed in collaboration with the National Insurance Crime Bureau (NICB).
The scale of the problem is hard to overstate. The Coalition Against Insurance Fraud's most recent study, published in 2022, estimates fraud costs the US economy $308.6 billion a year. North Carolina's Department of Insurance puts that at roughly 20 cents of every premium dollar – a tax ultimately paid by policyholders.
Gen Re said carriers carry both a responsibility and a legal obligation to shield clients from that fallout. The issue bites hardest in liability claims with a medical component, where examiners often stand as the first line of defense.
Settlement packages from plaintiff representatives can sometimes appear out of step with the underlying loss, with no treatment recorded at the scene yet months or years of subsequent care, multiple surgeries, extensive diagnostics and a sizable lost-wage component tied to permanent disability claims.
Examiners are urged to watch for undisclosed pre-existing conditions, falsified records and invoices for services not rendered. Common patterns include double billing, phantom billing and upcoding – billing a more expensive service than the one delivered.
Engaging a neutral physician for an independent medical exam can validate the cause and extent of injuries, though leaning too heavily on the same examiner can erode credibility. Bill audits help flag inflated charges or treatments at odds with reported injuries, while social media monitoring and surveillance can expose gaps between a claimant's stated limitations and observed behavior.
AI applications, including pattern recognition, predictive analytics and natural language processing, are increasingly used to surface inconsistencies earlier.
Gen Re flagged that while AI is reshaping detection, it has also armed a new generation of bad actors. Industry data backs that up. Synthetic voice fraud aimed at insurers jumped 475% in 2024, far steeper than banking's 149% rise, Pindrop figures cited by Bright Defense show.
Deloitte's Center for Financial Services has projected US fraud losses tied to generative AI will climb from $12.3 billion in 2023 to $40 billion by 2027.
The arms race is now well under way. Despite heavy investment in AI-powered detection, insurers remain locked in continuous combat with fraudsters. Verisk's Riedman has said carriers are "effectively up against an entire ecosystem of rapidly evolving technology," with intelligence-pooling becoming standard practice.
The NICB said the accessibility of generative AI tools has lowered the barrier to filing false claims. Activity that once required corrupt providers, forged signatures and physical document tampering can now be executed in minutes.
Criminals can spin up as many as 20,000 fake IDs at a time, sold online for as little as $5 each. Generative AI is also being used to fabricate police accident reports and medical records, while voice cloning produces personas convincing enough to handle live calls with insurance staff.
