AI Impact on Insurance Industries in Greater China
June 2025
Authors
Jian Gang He, FSA, FCIA
Hong Li, FSA, ACIA
Executive Summary
This report examines how artificial intelligence (AI)—particularly large language models (LLMs)—is transforming the insurance sector in China and related regions, with parallel insights from academia. We review the industry’s decade-long adoption of “narrow AI” (traditional machine learning and automation) and analyze recent developments since the emergence of ChatGPT through literature, case studies, and original survey data.
We present findings from two surveys conducted in early 2025: one of 21 Chinese insurance companies and another of 18 academic institutions with actuarial programs. These complementary perspectives reveal significant trends, challenges, and future directions for AI in insurance.
Key findings
Prior to ChatGPT, Chinese insurers widely applied AI in online sales, customer service chatbots, fraud detection, and automated underwriting. By 2021, nearly half were using AI in customer service and one-third in claims and sales channels. The release of ChatGPT in 2023 catalyzed a rapid expansion of LLM applications: over 60% of surveyed insurers now report having at least one LLM-based application in production, and all are actively exploring the technology.
Use cases are broad and growing. Internal process support is the most widely adopted application (95%), followed by customer service (67%), agent assistance (57%), underwriting (52%), and claims processing (52%). Chatbots and knowledge-based Q&A systems are dominant in customer-facing applications, with nearly all adopters reporting improvements in timeliness and efficiency. Meanwhile, insurers are beginning to apply LLMs in product development and customer analytics, though uptake in those areas remains more measured.
Implementation strategies are typically hybrid. A majority of insurers (71%) co-develop LLM solutions with technology providers, while 62% also pursue in-house development. Among internal users, the open-source DeepSeek model dominates, used by over 90% of self-building firms.
Leading insurers such as Ping An, PICC, China Life, and Sunshine Insurance have launched dedicated LLM initiatives, often in partnership with major tech platforms, aiming to reshape workflows across distribution, operations, and risk functions.
On the academic side, all surveyed institutions conducting LLM research focus on insurance applications, and most have integrated LLM-related content into their curricula—primarily via lectures, case studies, and increasingly, hands-on student projects.
Material
AI Impact on Insurance Industries in Greater China
Acknowledgments
The authors’ deepest gratitude goes to the volunteers who contributed to this study:
Project Oversight Group:
Sujeeva Fernando, ASA
Michael Levine, FSA
Yifan Zhang, FSA
Lin Zhang, Professor
Questions or Comments?
Give us your feedback! Take a short survey on this report. Take Survey
If you have comments or questions, please send an email to Research@soa.org