Unpredictable AI Risks
Some Overseas Insurers Move to Strengthen Exclusions
"Developing New Markets"...Strategies to Expand Coverage
The rapid spread of generative artificial intelligence (AI) is emerging as a new variable for the insurance industry. As liability risks stemming from errors in AI outputs grow, some overseas insurers are moving to narrow their coverage, while others view this as a new growth area and are expanding their product offerings.
Increasing AI-related liability risks are making insurers' underwriting decisions and loss ratio management more complex. ChatGPT-generated image
According to the insurance industry on the 24th, some overseas insurers are reviewing ways to incorporate an exclusion endorsement for errors related to generative AI, developed by the Insurance Services Office (ISO), a U.S. organization that drafts standard insurance policy forms, into their liability insurance policy wording. This move is seen as reflecting concerns that, if AI computational results are used without verification, it will be difficult to manage loss ratios in professional liability insurance (E&O), directors and officers liability insurance (D&O), and commercial general liability insurance.
In particular, experts point out that, in the medical and legal fields, citing AI-generated case-law analysis materials or diagnostic advice without sufficient verification could lead to disputes over false information or misdiagnosis. Because AI generates answers based on probabilities, it is difficult to completely eliminate errors. In addition, the internal computational process by which results are derived is not transparently disclosed, so even when an incident occurs, there are limits to pinpointing the starting point of liability. From an insurer's perspective, a new type of risk has emerged for which it is difficult to estimate both the probability of occurrence and the expected loss amount.
AI insurance market expands...AI-specific coverage emerges
Conversely, some insurers see AI risks as a new business opportunity. Insurtech company Amilla AI and People's Insurance Company of China (PICC), among others, are already offering products that specifically cover risks that may arise from the use of generative AI. AXA's U.S. subsidiary AXA XL has also introduced a cyber insurance extension endorsement that covers issues such as data distortion in the course of using AI and the potential for copyright disputes. On this, Sophie Farhan, Global Chief Underwriting Officer for Cyber at AXA XL, said, "AI is both a challenge and an opportunity for cyber insurance," adding, "Risk awareness and coverage structures related to AI are also evolving in line with these changes." Global consulting firm Deloitte projected that the global AI risk insurance market will continue to grow at a high compound annual rate of about 80 percent, reaching around 4.8 billion dollars by 2032.
In the case of domestic companies, the use of generative AI is still at an early stage, and it is assessed that full-scale development of related products has yet to begin. An industry official said, "The frequency of generative AI risks may be low, but once an incident occurs, the loss can be very large," and added, "Stable rate setting requires statistics on both the frequency of incidents and the scale of losses, but the lack of sufficient related data is a burden."
Han Jinhyun, a research fellow at the Korea Insurance Research Institute, also said, "There are not many cases of damage caused by the use of generative AI by domestic companies, and demand for related insurance products may not be very strong." He continued, "For the domestic insurance industry to provide coverage for generative AI risks in the future, it will first need to identify demand and then set coverage items so that they are compatible with the regulatory framework," adding, "In this process, it will be necessary to selectively introduce products starting with risks that have a higher likelihood of being underwritten."
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.


![Clutching a Stolen Dior Bag, Saying "I Hate Being Poor but Real"... The Grotesque Con of a "Human Knockoff" [Slate]](https://cwcontent.asiae.co.kr/asiaresize/183/2026021902243444107_1771435474.jpg)