China Launches National AI Roadmap for Pharmaceuticals, Medical Devices, & Cosmetics Supervision

On April 2, 2026, China’s National Medical Products Administration (NMPA) published the “Implementation Opinions on ‘Artificial Intelligence + Drug Regulation’” (“国药监综〔2026〕6号”), together with an official policy interpretation document. The framework establishes China’s long-term roadmap for integrating artificial intelligence into the supervision of drugs, medical devices, and cosmetics, with implementation milestones extending through 2030 and 2035.

The initiative represents a significant expansion of China’s smart-regulation strategy and reflects broader national priorities around healthcare modernization, digital governance, industrial upgrading, and AI-enabled public administration.

While many implementation measures remain high level, the framework provides an important strategic signal: China is moving toward a more connected, data-driven, and continuously supervised regulatory environment in which AI-supported systems increasingly underpin review, risk management, inspections, traceability, and lifecycle oversight across regulated healthcare products.

For pharmaceutical, biotech, medical device, and cosmetics companies operating in China, the framework is less about immediate new compliance obligations and more about the long-term direction of China’s regulatory architecture.

China Is Expanding AI-Enabled Lifecycle Regulation

The framework builds on “smart regulation” initiatives developed during China’s 14th Five-Year Plan period and aligns closely with broader healthcare and industrial modernization priorities emerging under the 15th Five-Year Plan and related State Council reform initiatives.

According to the NMPA, the objective is to establish a more integrated, intelligent, data-driven regulatory ecosystem using:

• Artificial intelligence
• Large language models
• Automation technologies
• Connected regulatory data systems
• Intelligent risk-monitoring capabilities.

The roadmap spans the full lifecycle of drugs, medical devices, and cosmetics, including:

• Clinical development
• Regulatory review and approval
• Manufacturing and distribution
• Traceability systems
• Inspections and enforcement
• Post-market monitoring and risk management

More importantly, the framework reflects points toward a broader shift from periodic and document-centric supervision toward increasingly connected, lifecycle-based, and continuously monitored regulatory oversight.

By 2030, the NMPA aims to establish:

• High-quality regulatory datasets
• Vertical large models dedicated to medical-products regulation
• AI agents supporting regulatory operations
• Human-machine collaborative review systems
• Intelligent risk-management and inspection capabilities

By 2035, the agency aims to establish a more integrated, data-driven, and intelligent regulatory governance system that is agile, collaborative, self-reliant, and operationally mature.

For companies operating in China, this signals continued movement toward more structured, digitized, and interoperable regulatory interactions extending beyond traditional electronic submission into lifecycle supervision, inspection readiness, traceability integration, and continuous risk monitoring. 

AI-Supported Review and Regulatory Operations Will Expand

The implementation opinions place significant emphasis on AI-supported regulatory operations across drugs, medical devices, and cosmetics, particularly in areas involving large-scale data processing, knowledge management, review support, and risk identification.

The NMPA plans to further standardize and structure regulatory data while expanding regulatory knowledge databases capable of supporting AI-assisted functions throughout review and supervision activities.

Planned AI-supported functions include:

• Product classification
• Task allocation
• Dossier review support
• Knowledge retrieval
• Risk and issue identification
• Report generation
• Certificate preparation and delivery

The policy specifically references development of AI large models and intelligent agents for “two products and one device” (“两品一械”), referring collectively to drugs, cosmetics, and medical devices.

Importantly, the framework does not suggest a transition toward autonomous regulatory approvals. Rather, it reflects the NMPA’s intention to use AI to improve consistency, efficiency, coordination, and regulatory responsiveness across increasingly large and complex regulatory data environments.

Provincial Authorities Will Accelerate Localized Digital Supervision

Provincial medical products administrations are expected to play a significant role in implementation. The NMPA instructed provincial authorities to accelerate intelligent applications for:

• Class II medical device review
• Post-approval drug change filings
• General cosmetics filing management
• Manufacturing and distribution licensing

This suggests that local review, filing, and supervision pathways are likely to become increasingly digitized and operationally standardized over time.

For companies operating across multiple provinces, this may gradually increase expectations around data consistency, submission structure, digital traceability, and system interoperability.

Human Accountability Remains Central

Despite the strong policy emphasis on automation and intelligent supervision, the NMPA repeatedly emphasized that AI systems are intended to remain “assistive” rather than fully autonomous.

The framework requires:

• Human review mechanisms
• Defined accountability structures
• Full audit trails
• Controlled deployment procedures
• Validation and assessment of AI models and algorithms

The agency also warned against fragmented system construction, duplicate platform development, and unreviewed deployment of AI applications.

F The framework positions AI primarily as a regulatory decision-support and operational-efficiency tool rather than a replacement for formal regulatory accountability. 

China Plans to Apply AI Across the Full Regulatory Lifecycle

The roadmap extends well beyond product review and approval. The agency plans to integrate AI-supported systems throughout clinical development, manufacturing oversight, distribution and use, post-market monitoring, inspections, enforcement, and public regulatory services.

This is particularly relevant for pharmaceutical and biotech companies, where regulatory oversight increasingly intersects with:

• Digital clinical systems
• Manufacturing-data environments
• Lifecycle traceability infrastructure
• Real-world safety monitoring
• Connected quality-management systems

Clinical Trial Data Governance Will Strengthen

Within the research and development stage, the NMPA plans to strengthen governance of clinical trial data and electronic systems.

The agency announced plans to develop supporting standards and technical guidance covering:

• Electronic clinical trial records
• Computerized system validation
• Broader clinical-trial data governance

The policy also encourages greater use of clinical trial big data to support regulatory oversight and supervision efficiency.

For pharmaceutical and biotech companies, this direction may gradually increase expectations around:

• Validated electronic systems
• Data integrity and auditability
• Structured clinical datasets
• Inspection-ready digital records
• Lifecycle data governance

This is particularly relevant for multinational sponsors operating global clinical programs that include China sites or rely on cross-border data environments.

Manufacturing Oversight Will Become More Data-Driven

The implementation opinions place particular attention on higher-risk product categories, including:

• Vaccines
• Blood products
• Special drugs
• Traditional Chinese medicine injections

The NMPA plans to deploy intelligent risk monitoring tools capable of utilizing:

• Production video streams
• Images
• IoT sensor data
• Real-time manufacturing information

The objective is to support more dynamic and risk-based quality supervision using both onsite and non-onsite oversight methods.

For pharmaceutical and biotech manufacturers, the framework points toward a broader regulatory direction emphasizing:

• More digitalized manufacturing oversight
• Integrated production and quality data management
• Enhanced lifecycle traceability
• More connected risk-monitoring capabilities

Over time, this may increase the strategic importance of inspection-ready digital evidence, data consistency, and integrated manufacturing governance frameworks.

Traceability System Will Continue Expanding

The NMPA also confirmed continued expansion of China’s pharmaceutical and medical device traceability system. The agency intends to accelerate traceability coding across all in-production product varieties and strengthen end-to-end traceability spanning production, distribution and product use.

The policy proposes development of integrated databases linking:

• Drug traceability codes
• Commercial barcodes
• Medical insurance codes
• Other relevant identifiers

China also plans to strengthen “trigger-based” supervision and enhance risk analytics using integrated traceability data.

For medical devices, the NMPA will continue promoting broader implementation of Unique Device Identification (UDI) systems across manufacturing, distribution, and post-market use.

Viewed strategically, these developments point toward increasingly connected lifecycle visibility across regulated healthcare products.

AI Will Support Risk Monitoring and Enforcement Modernization

A major component of the initiative involves AI-supported risk detection and enforcement modernization. The NMPA intends to establish systems capable of supporting:

• Multi-source risk signal aggregation
• Intelligent risk analysis
• Warning and early-warning functions
• Cross-regional coordination
• Complaint and monitoring-report analysis
• Online sales monitoring
• Public-opinion monitoring

Together, these initiatives suggest continued movement toward more connected and continuously monitored regulatory supervision models.

Intelligent Inspection Systems Will Expand

China also plans to modernize inspection and enforcement systems using AI and big data analytics. According to the policy, future inspection models may include:

• Risk-based inspection frequency
• Data-driven inspection planning
• Intelligent identification of compliance issues
• Mobile inspection tools
• Automated document and report drafting
• Scan-code entry for enterprise inspections

Provincial authorities are encouraged to establish unified inspection and enforcement platforms supporting city- and county-level regulators.

The framework therefore points toward more digitalized, data-driven, and remotely enabled inspection and enforcement models.

This may gradually increase the importance of maintaining structured, digitally traceable, and inspection-ready compliance evidence across manufacturing, quality, and post-market operations.

China Is Building Dedicated AI Governance Infrastructure

The implementation opinions also establish a broader governance framework supporting AI deployment within medical-products regulation.

Dedicated Regulatory Datasets Will Be Developed

The NMPA plans to establish high-quality regulatory datasets based on:

• Product dossiers
• Enterprise credit records
• Regulatory databases
• Laws and regulations
• Typical case databases
• Lifecycle regulatory data

The agency emphasized standardized data collection, annotation, and knowledge-extraction methods to support training and fine-tuning of regulatory AI models.

AI Governance and Cybersecurity Controls Will Tighten

The framework includes requirements covering:

• Algorithm transparency
• Model validation
• Cybersecurity monitoring
• Data security management
• Sensitive information protection
• AI risk assessment
• Infrastructure security

The NMPA specifically instructed regulators to prevent confidential or sensitive information from being entered into non-secure AI systems.

The agency also plans to establish governance mechanisms covering:

• Model-construction access controls
• Security review procedures
• Scenario compliance review
• Model and algorithm filing management
• Validation and assessment of AI applications

What Companies Should Monitor

The implementation opinions represent an important strategic step toward a more integrated, AI-enabled, and data-driven regulatory system across China’s healthcare-product sectors.

While many measures remain directional and further technical implementation guidance is expected over time, the broader trajectory is increasingly clear:

• More connected regulatory systems
• Greater reliance on structured and interoperable data
• Expanded lifecycle traceability
• Increasingly digitalized inspections and supervision
• More continuous and risk-based oversight models

Across all regulated sectors, companies should prepare for increasingly connected, traceable, and inspection-ready digital regulatory environments.

Viewed together, these developments suggest continued movement away from isolated and periodic compliance activities toward more integrated and continuously connected regulatory oversight.

Pharmaceutical & Biotech Companies

For pharmaceutical and biotech companies, this shift will likely be most visible in how clinical, manufacturing, traceability, and post-market data are increasingly connected and evaluated across a more integrated regulatory framework.

The overall direction of the policy suggests movement toward a more unified and digitally connected regulatory evidence framework spanning clinical development, regulatory review, manufacturing oversight, traceability, inspections, and post-market monitoring.

Regulatory readiness will increasingly depend not only on the quality of submission dossiers, but also on the consistency, interoperability, traceability, and auditability of lifecycle data across global and China operations.

Medical Device Companies

For medical device manufacturers, the focus is shifting toward tighter linkage between product identification, supply chain data, and real-world use.

In practice, the policy direction suggests increasingly data-connected oversight across the medical device lifecycle.

Cosmetics Companies

For cosmetics companies, some of the earliest operational changes may emerge in filing processes and ongoing supervision, particularly at the provincial level.

The direction is toward faster, more automated administrative processing paired with tighter ongoing digital oversight.

Final Thoughts

China’s AI-enabled regulatory framework marks a significant step in the country’s broader transition toward more connected, data-driven, and lifecycle-based supervision across drugs, medical devices, and cosmetics.  With implementation milestones extending through 2030 and 2035, the NMPA is building an increasingly integrated regulatory ecosystem in which AI supports product review, inspections and enforcement, traceability systems, manufacturing oversight, real-time risk monitoring and regulatory coordination and governance.

At the same time, many companies continue to operate across fragmented data environments, siloed systems, and legacy processes that were not originally designed for increasingly connected and AI-supported regulatory oversight. As regulatory expectations evolve toward greater interoperability, traceability, structured data governance, and digitally supported supervision, these gaps may increase:

• Regulatory follow-up
• Inspection complexity
• Data-management burden
• Operational inefficiencies

particularly in areas such as lifecycle traceability integration, manufacturing transparency, and inspection-ready digital evidence management.

While many implementation measures remain high-level and further technical guidance is expected over time, the strategic direction toward more connected, intelligent, and continuously monitored regulatory oversight is increasingly clear.

For global manufacturers, MAHs, and regulated companies operating in China, the key strategic question is no longer whether China’s regulatory system will become more digitalized, but how quickly organizations can adapt their governance models, operational infrastructure, and data environments to increasingly connected and AI-enabled supervision.

To help companies navigate these developments, Cisema provides strategic regulatory, compliance, market access, and operational support for pharmaceutical, medical device, and cosmetics companies across China and the broader Asia-Pacific region.

For companies seeking to turn these policy developments into practical, risk-based compliance strategies, contact Cisema today to learn how we can support your China market readiness.

Further Information

Explore Cisema’s services for:

• Pharmaceutical Consulting
• Discover Cisema’s medical device registration services in China
• Explore Cisema’s cosmetic product registration services in China

Read Cisema’s analysis of China’s AI ethics review and its implications for health technologies

References

• Implementation Opinions on “Artificial Intelligence + Drug Regulation” (NMPA, No. 6 of 2026)‍
• Policy Interpretation of the “Artificial Intelligence + Drug Regulation” Implementation Opinions