ZhongAn Insurance and ZhongAn Technology have released China’s first-ever white paper on generative artificial intelligence (AI) technology, which is predicted to emerge as an essential strategic asset for insurance companies in the future.
The ‘AIGC/ChatGPT Insurance Industry Application White Paper’ provides an in-depth analysis of more than 30 specific areas where AIGC technology could be implemented in the insurance field, including product design, actuarial science, marketing, operations, and customer service.
Additionally, they have the potential to optimise daily operations and improve research and development efficiency for insurance companies. All these along with ZhongAn’s own exploration of the technology’s potential.
The document also highlighted obstacles faced by the industry in utilising AIGC generative artificial intelligence technology, such as copyright hurdles during marketing, difficulties in adapting insurance data for model training, and potential risks concerning the confidentiality of customer data.
ZhongAn added: “In the operational realm, compliance issues surrounding the proper use of core insurance data and the accuracy of output results may come to the forefront. In practical application, enterprises must take a comprehensive approach to consider data quality, semantic comprehension, security risks, and user demands to ensure the model’s precision and usability.
“In conjunction, developers and users must possess the necessary skills and experience to blend artificial intelligence with human judgement to achieve optimal collaborative efficiency.”
This groundbreaking development, the insurer highlighted, marks a significant step forward in the integration of AI into the Chinese insurance sector.
Jiang Xing, the CEO of ZhongAn Insurance, said: “The inflection point in the evolution of artificial intelligence technology has arrived, and we firmly believe that its potential and speed of evolution in the future will far exceed our imagination.
“As stated in the report, the Transformer structure resolves the issue of parallel computing, enabling the model to train on a vast corpus within a reasonable timeframe. Consequently, AIGC has superior abilities to comprehend and interpret complex sentence structures, semantic relationships, and logical reasoning. It can also grasp long-term dependencies in text, comprehend context and meaning, and generate more accurate, coherent, and semantically enriched text.”
