China Unveils Plan to Build AI Metrology Standards
China has taken a major step toward standardizing artificial intelligence quality assurance, as two government departments jointly issued a comprehensive plan to build AI metrology capacity. The guidelines aim to address fundamental challenges hindering AI deployment in critical sectors by making AI performance “measurable, comparable, and traceable.”
The State Administration for Market Regulation (SAMR) and the National Development and Reform Commission (NDRC) released the “AI Metrology System and Capacity Building Guidelines (2026 Edition)” on May 28, according to Xinhua News. The document systematically lays out a framework for establishing measurement standards and quality assurance systems for AI technologies across industries.
Tackling the AI ‘Black Box’ and ‘Data Famine’
The guidelines zero in on two of the most persistent obstacles in AI development: the algorithm “black box” problem and the shortage of high-quality training data, often referred to as “data famine.”
Regarding the black box issue — where AI decision-making processes are opaque and difficult to interpret — the guidelines deploy key technology research for internal state monitoring and characterization of AI systems. The goal is to establish reliable, safe, and trustworthy AI metrology standards, as reported by CCTV News.
“Focusing on the ‘unmeasurable’ problem, making AI more trustworthy,” CCTV reported, the guidelines target pain points such as algorithm black boxes and poor decision interpretability, aiming to achieve AI technology performance that is “measurable, comparable, and traceable.”
On the data front, the guidelines call for constructing datasets with the highest metrological characteristics, standard reference datasets, and test datasets. A basic resource sharing mechanism will be established to break industry data barriers and enable secure data sharing, providing precise “ammunition” for AI algorithm training and evaluation.
A ‘Weights and Measures’ for the AI Era
The guidelines draw an explicit parallel to traditional metrology — the science of measurement that underpins all advanced economies. Just as standardized weights and measures were essential for trade and industry, the document envisions a unified measurement framework for AI products.
CCTV reported that the guidelines support building national-level metrology technology research and development centers and developing a batch of AI metrology standard devices with independent intellectual property rights. This will accelerate the formation of full-chain metrology capabilities covering algorithm models, computing efficiency, and data quality, providing a unified “weights and measures” for AI products.
The initiative aligns with China’s 15th Five-Year Plan (2026-2030), released in March 2026, which emphasizes comprehensive implementation of the “AI Plus” initiative.
14 Priority Sectors for AI Integration
The guidelines identify 14 key application sectors where metrology technology will be deeply integrated, including smart manufacturing, smart healthcare, and smart transportation. In each sector, the guidelines call for metrology research on critical parameters — such as the reliability of AI diagnostic algorithms in healthcare — to help solve quality assessment challenges in industrial digital transformation.
According to CGTN, the release of the guidelines marks “a critical shift in China’s AI sector from simply expanding computing power and scale to improving quality and strengthening foundational capabilities.”
From Scale to Quality: A Strategic Pivot
The guidelines signal a maturing approach to AI development in China. The country’s AI core industry scale exceeded 1.2 trillion yuan in 2025, with over 30% application penetration in large-scale manufacturing. Rather than continuing to focus primarily on expanding computing power and model scale, the new framework emphasizes quality assurance and foundational infrastructure.
This shift reflects a broader policy direction. The 2026 Government Work Report called for “creating a new form of intelligent economy,” and the concept of “new quality productive forces” — a key policy concept promoted by the Chinese government — is directly supported by the metrology initiative.
Next Steps and Implementation
SAMR has announced plans to build a number of AI metrology technology R&D and application centers, with pilot programs in smart regulation and smart healthcare. These pilots are expected to form replicable “AI + Metrology” application scenarios that can be scaled across industries.
The emphasis on developing metrology devices with independent intellectual property rights also reflects China’s broader push for technological self-reliance amid ongoing tech competition with the United States.
Global Context
China’s move parallels international efforts to establish AI standards and governance frameworks. The United States announced in May 2026 that it would assess new AI models before their release, while Germany set a fourfold AI capacity target for 2030. China’s approach to AI metrology could influence international standards, particularly as the country seeks to shape global AI governance frameworks through bodies such as ISO/IEC.
As the guidelines move toward implementation, the key question will be how China’s metrology framework interacts with emerging international standards — and whether this push for measurable, trustworthy AI can accelerate deployment in the critical sectors where AI adoption has been slowest.
— Reporting contributed by Xinhua News, CCTV News, and CGTN