In recent years, AI medical large models have become a new hotspot in the digital transformation of hospitals. Despite various hospitals launching AI models targeting areas such as cardiovascular, pathology, and urology, few are willing to invest substantial funds, with many institutions still in观望 mode.
On the market, leading tertiary hospitals such as Shanghai Zhongshan Hospital, Ruijin Hospital, and Renji Hospital have all publicly released AI models, but the funding behind these efforts largely depends on local government procurement projects. For instance, Changzhou No. 1 People's Hospital initiated two public tenders in the first half of this year, with a total budget close to 10 million RMB for purchasing an AI medical large model platform. Industry insiders point out that although AI medical models have shown application potential in vertical fields like pathology, they still face numerous challenges in large language model (LLM) applications.
Image source note: Image generated by AI, image authorization service provider Midjourney
With technological advancements, various medical AI all-in-one machines have emerged on the market. These devices integrate software, computing hardware, and security modules, with prices typically ranging from tens of thousands to hundreds of thousands of yuan. However, the actual usage effects of these all-in-ones have been less than satisfactory. Some hospitals found after purchase that their performance was insufficient, leading to idle equipment gathering dust.
Despite these issues, the potential of AI medical large models remains significant. For example, the pathology large model jointly developed by Ruijin Hospital and Huawei achieved good results in specific fields, with diagnostic accuracy approaching 70%. This indicates that with sufficient data integration and algorithm optimization, AI applications in medical scenarios can reach or even surpass human doctors' levels.
However, industry experts believe that before hospitals purchase these AI products, they need to see clear test results. Currently, most hospitals are still in the trial stage, which also provides opportunities for manufacturers to collaborate with top-tier hospitals to develop standardized services for smaller hospitals through joint research.
With the participation of local mobile operators, the application scenarios of AI medical large models will further expand. The successful deployment of DeepSeek "AI Computing All-in-One" at Tianjin Medical University General Hospital marks the gradual implementation of AI medical technology in practical applications.