China Releases First National Standard for Artificial Intelligence Medical Devices

On July 30, 2026, the State Administration for Market Regulation (SAMR) and the Standardization Administration of China (SAC) published GB/T 47941.1-2026, “Artificial Intelligence Medical Devices — Quality Requirements and Evaluation — Part 1: Terminology.” The standard represents China’s first national standard specifically addressing artificial intelligence (AI) medical devices.

Developed under the oversight of the National Medical Products Administration (NMPA) and taking effect on August 1, 2027, this standard represents China's first national framework specifically the development, testing, registration, and evaluation of AI medical devices and software as a medical device (SaMD). It replaces a patchwork of industry references, such as YY/T 1833.1, with a unified nationwide baseline.

With a 12-month transition window, the standard provides international and domestic manufacturers a crucial period to harmonize their technical documentation and registration submissions.

Understanding the Scope: What Does GB/T 47941.1-2026 Cover?

The significance of GB/T 47941.1-2026 lies in its breadth. Rather than addressing a single type of AI technology or medical device, the standard establishes terminology across five areas spanning AI technology, data, quality, algorithm evaluation, and clinical application.

This is particularly relevant for manufacturers of:

• AI-enabled medical imaging devices
• Computer-aided diagnosis software
• AI-assisted clinical decision-support technologies
• Software as a medical device (SaMD)
• AI-enabled in vitro diagnostic (IVD) products
• Medical devices incorporating machine-learning models

For international companies, terminology can be a practical regulatory issue. Global technical documentation may use concepts and definitions that do not map directly onto Chinese regulatory terminology. Differences can create inconsistencies between product specifications, testing reports, technical files, clinical documentation, and registration dossiers.

By establishing nationally recognized terminology, GB/T 47941.1-2026 provides a common reference for manufacturers, testing institutions, and regulatory reviewers.

The five areas covered by the standard are as follows:

1. AI Medical Device Technical Concepts

The standard defines concepts associated with artificial intelligence technologies used in medical devices, including terminology related to AI algorithms, machine learning models, software systems, and AI-enabled functions.

A common technical vocabulary can help manufacturers use terminology consistently in product specifications and regulatory documentation throughout the development and registration process.

2. AI Medical Device Data Sets

The standard provides terminology related to data sets used during AI development and evaluation, including training data, validation data, and test data.

This is particularly relevant to machine-learning-based medical devices, where the selection, use, and evaluation of data are closely linked to algorithm performance and product quality.

3. AI Medical Device Quality Characteristics

Another core area is the terminology used to describe AI software and algorithm quality. The standard addresses concepts including:

• Functional safety
• Accuracy
• Robustness
• Performance
• Reliability

For AI-enabled diagnostic and medical imaging technologies, these characteristics are central to demonstrating that a product performs as intended and can be evaluated consistently.

4. AI Algorithm Quality Evaluation

The standard establishes terminology used to evaluate the quality and reliability of AI algorithms. A consistent vocabulary can support communication among manufacturers, testing institutions, and regulatory reviewers, particularly as AI performance becomes an increasingly important component of medical device assessment.

5. Clinical Applications of AI Medical Devices

The terminology framework also addresses AI applications in clinical settings, including AI-assisted diagnosis, medical image analysis, intelligent screening, and other healthcare applications.

This provides standardized language for describing the intended use and functions of AI-enabled medical devices in regulatory and technical documentation.

The Road Ahead: Future Standards

While GB/T standards are technically recommended rather than mandatory, they function as vital technical benchmarks within China's medical device ecosystem. More importantly, Part 1 is only the entry point for an expansive series of national standards dedicated to AI medical device quality and evaluation.

Additional parts of the GB/T 47941 series are already under development and are expected to address:

• Data sets
• Data annotation
• Algorithm traceability
• Pre-trained AI models
• Synthetic data
• Installation verification

Because China's regulatory framework for AI medical devices is evolving, international manufacturers should treat this release as the foundation of a broader compliance trajectory rather than an isolated update.

How Should Overseas Manufacturers Prepare?

With a 12-month window before enforcement begins, companies planning to register or commercialize AI medical devices in China should execute a two-step preparation strategy.

Review Current and Pending Regulatory Documents

Review technical documentation, product specifications, user manuals, registration materials, and other regulatory files against the terminology established by GB/T 47941.1-2026.

Companies with products undergoing registration or being prepared for submission should determine whether updates are necessary, particularly for AI-enabled software, diagnostic systems, medical imaging technologies, and IVD products. Identifying differences early can help prevent inconsistencies during the China medical device registration process.

Align Quality and Technical Documentation

Manufacturers should consider incorporating standardized terminology into relevant quality management systems, technical files, clinical evaluation documents, and software lifecycle documents.

Consistent terminology across development, testing, registration, and post-market activities can support regulatory consistency throughout the product lifecycle.

Final Thoughts

GB/T 47941.1-2026 marks an important step in the standardization of AI medical devices in China. By establishing nationally recognized terminology, the standard creates a more consistent foundation for the development, testing, evaluation, and registration of AI-enabled medical technologies.

For overseas manufacturers, the immediate priority should be to understand how the new terminology applies to their products and regulatory documentation. The approximately 12-month transition period also provides an opportunity to review existing and planned China regulatory submissions.

International manufacturers should use the transition period to review relevant documentation and monitor the remaining standards in the series. To assess how GB/T 47941.1-2026 may affect your products and regulatory documentation, contact Cisema today.

Further Information

• Explore Cisema’s services for China NMPA medical device registration.
• Read more about AI medical device and SaMD classification in China.
• Explore Cisema’s overview of China’s digital health regulatory framework for SaMD (requires RAPS login).

References

• The national standard “Quality Requirements and Evaluation of Artificial Intelligence Medical Devices — Part 1: Terminology” was officially released. 《人工智能医疗器械 质量要求和评价 第1部分：术语》国家标准正式发布.
• Artificial Intelligence Medical Devices — Quality Requirements and Evaluation — Part 1: Terminology. 人工智能医疗器械 质量要求和评价 第1部分：术语.