(Yicai) Sept. 18 -- Large language models are expected to take small Chinese hospitals by storm, improving diagnosis and treatment, according to the vice president of a major clinical testing laboratory operator.
The most valuable point of applying artificial intelligence in medicine is LLMs in primary healthcare institutions or community clinics, Li Yinghua, VP of KingMed Diagnostics, said to Yicai. In July, Shanghai-listed Kingmed launched its open innovation platform for clinical testing and diagnosis which is deemed to cover vast amount of medical knowledge.
Many hospitals in China already use AI to assist treatment but LLMs help bring barrier-free natural language interactions between humans and machines, which can mobilize resources to solve problems in diagnosis and treatment, Li said, adding that he expects to see more treatments based on LLMs in the next year or two.
AI could act as a family doctor to make preliminary diagnoses for patients, recommend hospitals for further diagnosis and treatment, and help patients analyze test reports, Li said, adding that application scenarios are unlimited.
Most excellent clinicians and pathologists are concentrated in big hospitals in major cities so services based on artificial intelligence-generated content can play a big role in ensuring that 90 percent of people with serious diseases can receive good treatment without leaving their counties, Li said.
Several companies are developing medical LLMs. Tencent Holdings released many medical AI application products, including a medical LLM, earlier this month. Tencent has strong capabilities for logical reasoning in complex Chinese contexts and reliable task execution, per Zheng Yefeng, a distinguished scientist of the internet giant.
Moreover, Weimai Technology debuted CareGPT, China’s first LLM application in health management, last month. Meanwhile, Ping An Healthcare and Technology has been internally exploring the use of generative AI in assisted diagnosis and treatment since last year, and the results are satisfactory, Fang Weihao, chairman and chief executive officer of the Hong Kong-listed company, disclosed in August.
Editors: Tang Shihua, Emmi Laine