— A Brief Analysis on the List of Typical Application Scenarios of Artificial Intelligence for Drug Governance
1. Introduction and Background
On June 18, 2024, the National Medical Products Administration of People’s Republic of China issued the List of Typical Application Scenarios of Artificial Intelligence for Drug Governance (the “List”), presenting fifteen application scenarios that can play a leading demonstration role, possess characteristic of development potential, address pain points during the work, and tailor to more urgent needs.
It is noted from the beginning of the List that, the following guiding regulations adopted previously are playing a fundamental role in framing the List:
No. | Guiding Opinions | Issuing Governmental Authority | Issuance Date |
1 | The 14th Five-year Plan for National Economic and Social Development of the People’s Republic of China and the Outline of the long-range goals for 2035 | The National People’s Congress | March 12, 2021 |
2 | Development Plan on the New Generation of Artificial Intelligence | The State Council | July 8, 2017 |
3 | Guiding Opinions on Enhancing Scene Innovation and Promoting High-quality Economic Development with High-level Application of Artificial Intelligence | Ministry of Science and Technology, Ministry of Education, Ministry of Industry and Information Technology, Ministry of Transport, Ministry of Agriculture and Rural Affairs, National Health Commission | July 29, 2022 |
4 | Drug Governance Network Security and Information Construction “14th Five-Year Plan” | National Medical Products Administration | April 24, 2022 |
It is widely and notably recognized that application scenarios are pivotal when considering artificial intelligence (“AI”) empowering and supporting a variety of industrial sectors. According to the above-listed guided opinions and plans, issuance of a list describing various application scenarios have been underscored as a standardized approach, aiming to integrate AI into traditional industrial fields such as drug research, production and operation.
The Circular Regarding the Issuance of the List (the “Circular”), promulgated by the General Affairs Department of the National Medical Products Administration also sets the goal of the List as follows:
- promoting the research and exploration of AI-backed technology in the field of drug governance,
- enhancing the deep integration of AI and drug governance as the main line,
- standardizing and guiding drug regulatory authorities at all levels to carry out research and application of AI technology,
- guiding the focus of resources and promote the AI-enabled drug monitoring system, and
- providing reference and guidance for related research and application of other scientific research institutions, technology companies and pharmaceutical enterprises.
2. Specific Guidance on AI-powered Application Scenarios
In order to thoroughly address the current issues regarding AI-backed drug and medical device, the detailed application scenarios have been initially divided into four major categories, which further encompass several specific scenarios in each group. The below table shows headlines of those major categories along with scenarios elaborated in each type.
No. | Category | Application Scenarios |
1 | Admittance and Approval | Procedural review |
Auxiliary review and evaluation | ||
Batch arrangement and processing | ||
2 | Routine Supervision | Remote supervision |
Onsite supervision | ||
Auxiliary sampling for examination | ||
Auxiliary inspection and case handling | ||
Drug alert | ||
Network transaction governance | ||
3 | Public Service | Service processing and policy consulting |
Elderly-oriented modification on instruction | ||
4 | Auxiliary Decision-making | Service data query |
Data analysis and forecast | ||
Work scheme inspection | ||
Risk management |
Pursuant to the List, the category regarding admittance and approval is related to threshold of entering the circle, which is likely to draw much of market players’ attention. In this regard, however, the List provides that we, people use AI technology to build a large language model (the “LLM”) based on relevant laws and regulations, in order to (i) realize the automatic intelligent review of the electronic application materials for the registration of drugs and medical devices, (ii) expediently determine the compliance of application materials, and (iii) analyze and compare the research data of the declared products, which in turn determine the authenticity of data in question at a preliminary phase, and then provide the specific basis regarding non-compliance.[1] With more AI embraced in the preliminary review and screening of application material, it merits further attention that pertinent applicants shall draft and submit the material in accordance with the instructions provided by the National Medical Products Administration so as to pass the AI review and screening model smoothly. Though the List, namely, is mainly focusing on drug governance, procedural review on medical device and cosmetics are also covered in this part, which could enhance efficiency and productivity in the relevant process.[2]
What’s more, worthy of mention is the emphasis of AI-powered LLMs in the List.[3] It is well-noted that the LLM has emerged as a bright spot for today’s AI advancements. The list actively responds to this development trend by directly referring to the application of LLM for several times, encouraging people to harness the LLM capacity in aspects of information searching and analyzing, generating initial draft of certain types of documents, data processing and forecast, and even the auxiliary review work, aiming to improve the working quality and efficiency, thereby boosting the high-quality development.
Last but not the least, the List makes clear the auxiliary aiding function of AI-related technology served in pertinent work, indicating roles and continuous efforts of staff involved like inspectors on such matter. For instance, with respect to onsite supervision and sampling for examination, certain work will still be primarily conducted by inspectors or reviewers.[4] This is not hard to believe since those work may require substantial discretion to occasionally decide on an ad hoc basis whether a particular activity should be penalized or excused under the routine scrutiny and surveillance. Similarly, during the decision-making process, AI’s aiding role is determined while individuals involved are highly encouraged to harness its strong computing capacity and data process abilities, even generating the initial draft of certain analysis and inspection reports.[5]
3. Guiding List and Open-ended Approach
AI-powered technology has become increasingly important in the development of pharmaceutical industry. For instance, AI-backed drug discovery has experienced boom in recent years. Meanwhile, it goes without saying that the pharmaceutical industry holds significant importance in terms of public well-being, people’s health conditions and human’s lifespan. In order to fully take advantage of the cutting-edge technology, maximizing its benefits brought to key industrial sectors, the guiding opinions, regulations or list in question can catalyze the process, facilitating the actual application in particular scenarios.
Admittedly, the network security and data security also merit emphasis. The Circular underlines that all related units shall pay attention to issues associated with the network security and data security. According to the category and classification-based protection requirements of data resources subject to governance rules and the computing capacity requirements required by the AI model, appropriate application deployment schemes shall be selected. In the meantime, system and data access rights shall be reasonably set to avoid the risk of data leakage and abuse, so as to ensure the safe and steady application and development of AI technology concerning drug governance, which has been fairly elaborated in existing laws and regulations like Network Security Law, Data Security Law and Personal Information Protection Law.
In addition, it is worth noting that the List, pioneered by the Chinese pharmaceutical authority, is the first guiding regulation on the national level that navigates the AI application scenarios in a vertical industry. Prior to issuance of the List, some governmental authorities in local level have delved into the discussion regarding application scenarios of AI or similar digital technology, granting some pilot projects in this regard.[6] Those local-level attempts mainly focus on the implementation while the List is crafted to offer higher-level guidance covering the pharmaceutical industry.
Due to its attribution, relevant policy makers have strived to employ an approach combining a typical specific list with a comparatively broad guideline. Certainly, market players involved shall comply with the existing legal framework in respect of network and data security as a whole. But the List concentrates on offering instructions empowering AI-backed technology in various subdivision of drug governance, leading to increasing application in practice. Undoubtedly, the List develops a sensible legal and policy response to the actual application of AI by listing possible scenarios. The implementation of the List along with realization of the aspirations implied in the Circular are well worth expecting and hope to open more opportunities in a variety of industrial sectors.
[1] See Application Scenarios No. 1 of the List.
[2] See supra note 1.
[3] See Application Scenarios No. 1, 2, 7, 9, 10, 11, 12, 13 and 14 of the List.
[4] See Application Scenarios No. 5 and No. 6 of the List.
[5] See Application Scenarios No. 9 and No. 13 of the List.
[6] According to the First Batch and the Second Batch of Key Construction Projects on Artificial Intelligence Application Scenarios in Hangzhou, issued by Hangzhou Municipal Bureau of Economy and Informatization on April 25, 2021 and June 29, 2021, respectively, specific projects were listed to facilitate the AI application scenarios in certain type of industries, such as fintech, smart healthcare, smart education, and so forth.