AI-Generated Forgeries Push Insurance Fraud to Record Levels
Insurance fraud in South Korea reached 1.16 trillion won ($788 million) last year, with authorities and insurers pointing to AI-powered document forgery as a major driver of the surge. The Financial Supervisory Service released the figures Tuesday, marking an all-time high for detected fraud.
Falsified medical certificates and inflated hospital claims account for 201 billion won of that total. Combined with staged car accidents, these three categories represent 35 percent of all detected fraud.
While officials have no separate statistics isolating AI-driven cases, the insurance industry sees artificial intelligence as the primary factor behind the increase. Forgers now use AI to fabricate hospital seals, doctor signatures, and entire document formats-techniques that are difficult to catch in the low-resolution images that dominate mobile-based claims.
Real Cases Show Detection Gaps
A woman in her 40s filed 39 claims over three years, citing conditions from finger fractures to cancer. When an insurer cross-checked submissions against hospital records, it discovered that documents, seals, and signatures had been created with AI. She obtained 25 million won before detection.
Last year also saw the first prison sentence for AI-based insurance fraud. A man in his 20s received a two-year sentence for using AI to manipulate diagnoses and extend hospital stays, defrauding insurers of 150 million won across 11 cases.
Mobile Claims Create New Vulnerabilities
The shift to photo-based claims has inadvertently made fraud easier. Slightly blurred, AI-altered images are harder to verify than physical documents, and smaller claims often escape scrutiny entirely.
"The ease of filing claims by simply taking a photo has unintentionally opened the door to more forgery," an industry official said. "For smaller claims, insurers may overlook inconsistencies, as detecting sophisticated AI edits in low-resolution images remains technically difficult."
Insurers Deploy AI Defenses
Insurers have begun using AI data analysis to flag irregularities in claim histories, medical records, and accident patterns. However, detecting advanced image manipulation remains a significant technical challenge, even with these systems in place.
Financial authorities acknowledge the problem. "Recent insurance fraud has become so sophisticated through the use of AI that it is difficult to detect not only with the naked eye but also with conventional screening systems," an FSS official said. "We are jointly developing advanced screening technologies with the insurance industry to identify such cases, but it is expected to take some time."
Direct Data Transfer Offers Long-Term Solution
Industry insiders say the most viable defense is expanding fully digital claims platforms where medical data flows directly from hospitals to insurers. This approach would virtually eliminate opportunities for policyholders to tamper with documents before submission.
For now, insurers remain caught between the speed and convenience of mobile claims and the security risks those systems introduce. AI for insurance applications continue to evolve on both sides of the fraud detection equation.
Your membership also unlocks:
