On 2 April 2026 the National Medical Products Administration (NMPA) issued the Implementation Opinions on “Artificial Intelligence Plus Drug Regulation” (Guo Yao Jian Zong [2026] No. 6), implementing the State Council’s opinions on deepening the “AI Plus” initiative and the reform of drug and medical-device regulation. The stated milestone is to establish an initial system integrating drug regulation with AI by 2030, with the operating and management mechanism for “AI plus drug regulation” essentially in place.
Why the regulator moves first
Most of this round of “AI Plus” implementation is happening on the industry side — manufacturing, consumption, transport, medicine. The NMPA’s decision to start from the regulatory side is worth noting.
Regulatory digitalisation applies directly to approval, inspection, monitoring and enforcement. Its value is not that regulation becomes faster, but that the data definitions, risk-identification standards and evidentiary forms on which regulation relies change. Once regulators use models to assist review and risk warning, the requirements on the structure and traceability of the material companies submit rise as well. In other words, capacity gains on the regulatory side translate into a higher compliance bar for companies.
Seven priority directions
Taking modernised drug regulation as the goal, the Opinions set out seven priority directions for regulatory digitalisation, covering the full regulatory chain of review and approval, inspection and enforcement, and monitoring and evaluation. Their common feature is embedding AI into existing regulatory processes rather than building a separate system.
For pharmaceutical companies, this means the way they interact with the regulator will gradually standardise: the format and quality requirements for submissions, clinical-trial data and pharmacovigilance information become clearer, and the room for “supplement and resubmit” narrows.
Five foundational tasks
Responding to new trends in AI technology, the Opinions set out five foundational tasks for “AI plus drug regulation”, covering data resources, computing facilities, standards, security and talent.
The value of these foundations lies in data. Regulatory digitalisation depends heavily on high-quality, usable industry data resources, which aligns with the National Data Administration’s push on public-data authorised operation and the building of high-quality industry datasets — see our analysis of the action plan for high-quality industrial datasets. For the data and AI service providers involved, lawful data sources, personal-information boundaries and security measures are preconditions — see our data-compliance service.
What it means in practice
First, companies building tools for the regulatory side face stricter entry requirements. For AI-assisted review, image-assisted diagnosis and similar functions provided to regulators or medical institutions, the compliance attributes of the algorithms and models, the data sources and explainability will be focal points in assessment.
Second, pharmaceutical companies’ own AI applications must be brought into a compliance framework. Drug discovery, real-world studies and patient services involve large volumes of personal and sensitive data, requiring a review of processing activities and authorisation chains through our data-compliance service.
Third, products offering generative functions to the public must still perform algorithm filing and synthetic-content labelling — see our algorithm filing guide.
Regulatory digitalisation is a long-term theme. Rather than waiting for each rule to land, companies are better off getting their data ledgers and compliance documents in order now — the foundation any version of the rules will require. For the wider governance approach, see our AI law practice.
This article is based on public information concerning the NMPA’s Implementation Opinions on “Artificial Intelligence Plus Drug Regulation” (Guo Yao Jian Zong [2026] No. 6) issued on 2 April 2026. It is for general reference only and does not constitute legal advice.