With breakthroughs in artificial intelligence (“AI”), the healthcare industry has experienced rapid development. Since 2016, China has issued a series of relevant policies, including the “Internet+” Artificial Intelligence Three-Year Action and Implementation Plan, Guiding Opinions on Promoting and Regulating the Application and Development of Big Data in Health and Medical Care and the “Healthy China 2030” Blueprint. These policies encourage and guide the integration of AI, such as natural language processing and machine learning, with various aspects of the medical and healthcare industry.
As medical AI applications continue to be commercialized, leading to significant cost reductions and enhanced efficiency, corresponding practical regulations have been gradually established.
In addition to general rules concerning personal data protection and data security, in July 2021, the Center for Medical Device Evaluation (CMDE) of the National Medical Products Administration (NMPA) issued the Guidance for the Classification Defining of AI-Based Medical Software Products. This document classifies and provides registration, filing, and clinical evaluation requirements for AI medical software products.
6 months later, the CMDE of the NMPA also issued the Guiding Principles for the Registration Review of Artificial Intelligence Medical Devices. These guiding principles classify AI medical devices by their software security level. Based on the expected use, usage scenarios, and core functions of the product, the software security level is categorised into mild, moderate and severe. Regulatory requirements for the AI medical devices will vary depending on the software security level. The guiding principles also offer comprehensive directives on several key aspects, such as the labelling, transfer learning, third-party databases, whitebox algorithms, pressure testing, adversarial testing, AI chips, and more. Significant attention is given to technical evaluations, system inspections, monitoring of adverse events, and the establishment of product recall requirements to ensure robust quality control throughout the entire life cycle of AI medical devices.
AI has also significantly impacted healthcare service delivery, giving rise to the popular concepts of “internet medical” and “remote medical”. In the realm of medical diagnosis within healthcare services, AI-based internet medical and remote medical practices leverage online platforms and software networks to facilitate hospitals and physicians in delivering convenient and efficient healthcare services to patients. These AI applications enable patients to receive prompt feedback, while also enhancing the diagnostic efficiency of medical professionals.
Beijing has recently taken lead to impose restrictions on the use of generative AI in online healthcare activities , although the regulations are still in the drafting phase and have not been finalized. Personnel who use AI medical devices will likely require professional qualifications. All physicians must undergo real-name authentication to ensure that the medical services are provided by the person themselves. AI medical software is not permitted to replace doctors in the delivery of medical diagnosis and treatment services.
Given the rapid development of AI and the dynamic nature of related regulations, it is crucial to remain attentive to these evolving requirements and guidelines.