On April 2, 2026, the National Medical Products Administration (NMPA) issued the Implementation Opinions on "Artificial Intelligence + Drug Regulation" (hereinafter referred to as the "Opinions").

The Opinions consist of four parts: General Requirements, Key Regulatory Scenarios Empowered by Digital Intelligence, Basic Support for "AI + Drug Regulation", and Organization and Implementation. Focusing on the key tasks of drug regulatory reform and aiming for the modernization of drug regulation, the document puts forward seven key directions for the digital intelligence of regulation in the next stage. In light of the new trends in AI technology development, it also proposes five key tasks to consolidate the basic support for "AI + Drug Regulation".

The Opinions set clear goals:

• By 2030, an innovation system integrating drug regulation and artificial intelligence will be initially established; the operation and management mechanism for "AI + Drug Regulation" will be basically formed; the computing power support base will be more intensive and efficient; high-quality datasets, vertical large models, and intelligent agents meeting the needs of intelligent regulation will be developed; AI will be effectively applied in scenarios such as review and approval, supervision and inspection, testing and monitoring, and government services; the efficiency of human-machine collaboration will be significantly improved; and the digital intelligent regulatory capacity throughout the product life cycle will reach a new level.

• By 2035, a new pattern of intelligent drug safety governance featuring data-driven, intelligent and agile, independent and controllable, and ecologically collaborative will be basically formed.

The release of the Opinions is an important measure for the NMPA to implement the decisions and arrangements of the CPC Central Committee and the State Council on comprehensively deepening reform and in-depth implementation of the "AI +" initiative. It will effectively promote the innovative application of artificial intelligence in the whole life cycle regulation of drugs, medical devices, and cosmetics, injecting strong impetus into ensuring public drug safety and promoting high-quality industrial development.

Implementation Opinions of the National Medical Products Administration on "Artificial Intelligence + Drug Regulation"

Document No.: Guo Yao Jian Zong [2026] No. 6

To the drug regulatory administrations of all provinces, autonomous regions, municipalities directly under the Central Government, and the Xinjiang Production and Construction Corps, all departments and directly affiliated institutions of the NMPA:

Since the implementation of the National Drug Smart Regulation Action Plan, drug regulatory authorities at all levels have actively explored the use of information technology to enhance regulatory capacity, and initially built a national integrated drug smart regulation system. Currently, the rapid development and iterative evolution of new-generation information technologies such as artificial intelligence provide new means and inject new momentum into smart regulation. To implement the Opinions of the State Council on In-depth Implementation of the "AI +" Initiative and the Opinions of the General Office of the State Council on Comprehensively Deepening the Reform of Drug and Medical Device Regulation to Promote High-quality Development of the Pharmaceutical Industry, seize the major strategic opportunity of AI development, promote the deep integration of artificial intelligence and drug regulation, and accelerate the modernization of drug regulation, the following opinions are hereby formulated.

I. General Requirements

Guided by Xi Jinping Thought on Socialism with Chinese Characteristics for a New Era, thoroughly implementing the spirit of the 20th CPC National Congress and its plenary sessions, fully implementing General Secretary Xi Jinping's important instructions and directives on drug regulation, adhering to the principle of leading drug regulation modernization with information technology, adhering to problem-oriented and systematic thinking, coordinating development and security, giving full play to the role of the smart regulation platform as a central hub, strengthening system coordination and open sharing, driven by data elements and guided by scenario applications, we will deeply promote the innovative application of artificial intelligence in the whole life cycle regulation of drugs. We will enhance the level of "one-stop online services, unified online governance, and online collaboration" through automation, precision, coordination, and intelligence, build a high-level national integrated drug smart regulation system, and provide strong digital intelligence support for comprehensively deepening drug regulatory reform.

II. Focusing on Key Scenarios of Smart Regulation, Empowering Regulatory Reform with Digital Intelligence

(1) Building a Human-Machine Collaborative Intelligent Review and Approval System

Promote the standardization and structuring of electronic submission of application materials, improve the review and approval knowledge base, accelerate the R&D and application of large models and intelligent agents for the review and approval of drugs, medical devices, and cosmetics, and efficiently empower scenarios such as intelligent product classification, task allocation, document review, knowledge retrieval, problem identification, report generation, and certificate issuance and delivery, significantly improving the quality and efficiency of review and approval.

Focusing on high-frequency scenarios in the review and approval work of local regulatory authorities, under the guidance of the NMPA and in accordance with the division of labor of provincial drug regulatory authorities, we will accelerate the implementation of intelligent applications in key scenarios such as the review of Class II medical devices, post-marketing change filing of drugs, filing of ordinary cosmetics, and production and operation license approval in a point-to-area manner, strengthen the transformation and sharing of achievements, and avoid low-level redundant construction.

Further improve the working system for AI-assisted review and approval, take ensuring the safety and effectiveness of products as the bottom line and improving the quality and efficiency of review and approval as the focus, establish and improve a human-machine collaboration mechanism of "digital intelligence empowerment, manual review, and full traceability", and accelerate the construction of an efficient, safe, and controllable intelligent review and approval system.

(2) Enhancing the Whole-Chain Intelligent Regulatory Capacity

Improve the digital intelligence level of the whole process. Accelerate the research and formulation of guiding principles for the standardized application of artificial intelligence in the pharmaceutical industry to adapt to the development needs of new industrial technologies. Promote the digital intelligence of the whole production and testing process of high-risk varieties such as blood products and traditional Chinese medicine injections, research and formulate supporting regulatory requirements, and gradually expand to other varieties, guiding the industry to improve the whole-process quality control capacity in accordance with regulations.

III. Grasping New Trends in AI Development, Consolidating Basic Support for "AI + Drug Regulation"

(1) Promoting the Construction of High-Quality Datasets for Drug Regulation

Adhere to the principle of "scenario-driven, priority for urgent needs". Centering on the core business scenarios of drug life cycle regulation and the actual needs of AI applications, we will promote the construction of high-quality datasets for drug regulation in stages and steps.

Further improve the national integrated drug regulatory data resource system, with national and provincial data centers as hubs, and based on product archives, enterprise credit archives, laws and regulations databases, and typical case libraries, improve the data aggregation and governance system, enhance data accuracy, consistency, and usability, and provide basic support for the construction of high-quality datasets.

Focusing on the training, fine-tuning, and implementation of vertical large models for drug regulation, clarify data format, quality, and content requirements according to scenarios, formulate scientific and unified collection specifications and annotation guidelines, carry out multi-source data fusion governance, professional annotation, and knowledge extraction, build general and professional knowledge bases in the drug regulatory field, and form high-quality datasets that are hierarchical, dynamically updated, and traceable throughout the life cycle.

On the premise of strictly ensuring security and privacy, orderly promote the compliant and efficient application of knowledge bases and high-quality datasets in scenarios such as model training, knowledge reasoning, and auxiliary decision-making.

(2) Strengthening the AI Application Support System

Adhere to business leadership, and coordinate the training, deployment, and application of large models in the drug regulatory field. Relying on existing smart regulation infrastructure, build a large model application and algorithm management platform, formulate model application guidelines and safety specifications, promote the co-construction and sharing of common technical components, improve model and algorithm management capabilities, and promote technology interoperability, resource sharing, and ecological collaboration.

Promote the deep integration of artificial intelligence and business information systems, and accelerate the large-scale implementation of AI-assisted regulatory scenarios. Centering on the whole life cycle regulation of drugs, build a multi-agent collaboration mechanism, improve the system linkage and business collaboration system, and promote the intelligent upgrading of drug regulatory capacity.

(3) Strengthening the Construction of Computing Power Infrastructure

The NMPA will make overall planning for a collaborative system of multi-level intelligent computing power resources, and national and provincial regulatory authorities will promote the supply of intelligent computing resources as needed. Build a standardized and scalable intelligent computing power base to meet the needs of intelligent applications in different network domains such as the Internet, government affairs extranet, and government affairs intranet.

Improve cross-domain collaboration and disaster recovery capabilities, gradually form a deployment pattern of "co-construction, co-governance, and sharing", enhance computing power support capabilities, and provide continuous and stable guarantee for regulatory intelligence.

(4) Consolidating the Security Protection System

Strictly implement the security responsibility system, upgrade the network security protection system, improve the mechanisms of network security situation awareness, information sharing, joint research and judgment, threat early warning, and traceability, use artificial intelligence technology to enhance the active network security protection capability, and build an intelligent and collaborative protection system.

Establish and improve the data security management system, clarify the catalog of core and important data, and improve the data security protection technology system. Strengthen AI risk monitoring and assessment, formulate algorithm transparency requirements and model verification specifications, strengthen the security capacity building of model algorithms, data resources, infrastructure, and application systems, strengthen AI application risk assessment and monitoring and disposal, prevent the input of classified and sensitive information into non-classified models, and promote compliant, transparent, and reliable AI applications.

(5) Improving the Construction and Operation Management Mechanism

Adhere to the auxiliary positioning of artificial intelligence in the field of drug regulation, clarify the functional boundaries and responsible subjects of large models and various intelligent auxiliary applications, and avoid behaviors such as deployment without review, multi-head construction, and redundant construction.

Establish a special mechanism to be responsible for the governance of AI applications in drug regulation, coordinate model construction access, security review, and scenario compliance review, formulate management systems for "AI + Drug Regulation", and clarify division of responsibilities and work norms.

Improve the model and algorithm filing management system, formulate basic guidelines and technical specifications, and conduct effectiveness and reliability verification and evaluation of models and their auxiliary applications. Strengthen the management of data resources such as training data, fine-tuning data, and knowledge bases to ensure legal sources, accurate content, compliant use, and full traceability.

Explore the authorized operation of public data for drug regulation, build a public data zone, promote authorization by field and according to scenarios, and strengthen the development and utilization of public data for drug regulation.

IV. Organization and Implementation

Drug regulatory authorities at all levels should deeply understand the new trends in AI development, regard it as an important starting point for supporting the comprehensive deepening of drug regulatory reform and a powerful support for enhancing drug regulatory capacity, coordinate and connect relevant plans, increase investment, promote the application of artificial intelligence in frontline regulation, promote construction through application, combine construction and application, and truly give full play to the effectiveness of artificial intelligence in regulation.

Strengthen demonstration and guidance, focus on the difficulties and bottlenecks of regulatory business, deepen the innovative application of smart regulation, and effectively empower business innovation. Strengthen the scientific and technological support of regulatory science research for "AI + Drug Regulation", and promote the implementation and transformation of relevant major scientific and technological projects.

Increase training efforts to improve the digital thinking, digital skills, and digital literacy of the cadre team.

National Medical Products AdministrationMarch 11, 2026

Source: National Medical Products AdministrationReposted by: Guangdong Drug Administration