The South Korean government is preparing a coordinated response to insurance fraud, forming a multi-agency task force to develop an AI-based detection platform as annual losses from confirmed fraudulent claims continue to climb.
The Financial Services Commission (FSC) convened the opening meeting of the “AI-Based Insurance Fraud Prevention System Task Force” and the “Insurance Investigation Council” on June 4, 2026. The gathering brought together a cross-section of public and private stakeholders, including the National Police Agency, the Financial Supervisory Service, the National Health Insurance Service, the Health Insurance Review & Assessment Service, the Korea Life Insurance Association, the General Insurance Association of Korea, and the Korea Insurance Research Institute.
Data published by the Financial Supervisory Service (FSS) and reported by the Korea JoongAng Daily on March 31, 2026, shows that insurers paid out 1.16 trillion won on fraudulent claims in 2025 – the highest figure on record and a 6.9 billion won increase from the previous year. The trend has been consistent over recent years, with fraudulent payouts standing at 1.08 trillion won in 2022 and rising to 1.12 trillion won in 2023. The number of identified suspects, however, moved in the opposite direction, declining 3% year-on-year to roughly 105,700 in 2025, according to the same report.
What regulators find more concerning is the gap between detected and undetected fraud. According to the Asia Business Daily’s report on the task force launch, the FSC puts the total estimated exposure – including cases that go undetected – at around 9 trillion won, a figure nearly eight times larger than confirmed payouts. Long-term non-life insurance products, a category that covers indemnity health and general health plans, represented the largest share of detected fraud at 44.7%. Auto insurance followed at 22.4%, life insurance at 21.8%, and general non-life at 11.2%. The FSC has framed the issue in terms that extend beyond the insurance sector. “Such leakage of insurance payouts leads to higher premiums, and when insurance fraud involves claims for health insurance benefits, concerns rise about the depletion of the health insurance fund,” an FSC official said, as reported by the Asia Business Daily.
The task force’s initial assessment points to a fraud landscape that is changing faster than existing detection systems can accommodate. Criminal networks involving medical institutions, auto repair shops, brokers, and agents are growing in scale and coordination, and technology is giving them new tools. Generative AI and deepfake technologies have lowered the barrier to document forgery. Using a smartphone alone, individuals can alter identification cards, medical certificates, and vehicle damage photographs at any point in the insurance process – from policy enrolment to claims settlement. When medical records or receipts are manipulated at the pixel level through AI generation, the font inconsistencies and spacing irregularities that human reviewers or basic software might catch simply do not appear. Additionally, repeatedly printing, outputting, and re-photographing a digitally altered file makes tampering progressively harder to identify.
Regulators identified an institutional gap that compounds the detection problem. Information sharing between financial sector bodies – such as the Korea Credit Information Services, the Korea Insurance Development Institute, and private insurers – and social welfare agencies like the National Health Insurance Service and the Health Insurance Review & Assessment Service remains limited. The infrastructure needed to cross-verify data held by these agencies, which would be central to detecting AI-assisted forgery, has not yet been built out.
To address these gaps, the government organized the TF around three subcommittees covering legal and institutional issues, data, and infrastructure, the Asia Business Daily reported. The division is intended to allow parallel progress across the regulatory, technical, and operational dimensions of the problem. The subcommittees are tasked with working through a set of core issues:
The TF has also established that policyholder protections – including privacy – will not be compromised in the process. Conventional fraud detection techniques, such as comparing submitted documents against originals, will continue alongside AI-based methods.
The FSC has set a September 2026 deadline for completing the platform development plan, with the task force expected to operate through that window. From October onward, the government intends to begin legislative and technical follow-through, including amendments to relevant laws and upgrades to the underlying platform. Kim Jinhong, Director General of the Financial Industry Bureau, outlined what a functioning system would mean in practice: “Once the AI-based insurance fraud prevention system is established and fully utilized, it will enable comprehensive response measures such as pre-emptive prevention, real-time detection, and post-fraud action, thereby reducing insurance fraud across the board. This is expected to enhance trust in the insurance industry, ultimately reduce premiums, and prevent leakage from the health insurance fund, delivering benefits back to the public.”
