Thread Content
According to Sinochem New Network, recently the group standard \"Guidelines for Artificial Intelligence Data Governance in the New Materials Sector\", which was drafted under the leadership of the Shenzhen New Materials Industry Association by 27 enterprises and research institutions including China Electronics Cloud, was officially released and put into effect. As the first dedicated standard in this field in China, it aims to address the key challenges related to data governance in the industry and provide standardized guidance for the intelligent transformation of the sector. With the deep integration of AI technology across the entire value chain of new materials, data has become a key production factor. However, the industry faces various challenges such as inconsistent data formats, variable data quality, inadequate safeguards for data security and privacy, and poor interoperability across different applications, all of which hinder the effectiveness of AI models and their practical implementation on a commercial scale. To address these challenges, the Shenzhen New Materials Industry Association brought together various stakeholders in the industry chain, including China Electronics Cloud, Betray New Materials Group, BYD New Materials, Sunwoda Electronics, and The Chinese University of Hong Kong, Shenzhen. After multiple rounds of research and evaluation, this standard was finally developed. It strictly adheres to GB/T 1.1-2020 \"Guidelines for standardization work\" and incorporates cutting-edge concepts such as materials genomics. As one of the main organizations involved in drafting this standard, China Electronic Cloud played a significant role in its development. Its practical experience and technical expertise in this field contributed to ensuring the standard’s scientific validity and practical applicability. According to him, the standard core content consists of four aspects. First, establish governance principles. Eight principles, including standardization, transparency, compliance, and security, were defined, with an emphasis on the traceability of experiments. Second is covering the entire process stage. A four-phase governance framework has been established, covering top-level design, organizational support, project construction, and operational optimization. Third, focus on multimodal data. For various types of data such as experimental characterization, simulation, manufacturing processes, and performance in applications, unified standards for collection, storage, and sharing have been established, with efforts being made to create cross-modal data linkage standards in order to break down data silos. Fourth, strengthen quality and safety. Specific data quality metrics have been defined, and security mechanisms such as data classification and grading, hierarchical encryption, and regional access control have been established; in addition, technologies like differential privacy and homomorphic encryption are used to safeguard privacy. This standard applies to various sectors such as alloys, new energy materials, semiconductors, and biopharmaceuticals, covering the entire process from research and development to industrial application. Its implementation is expected to provide enterprises with standardized operation manuals to reduce costs and improve efficiency, help overcome barriers to data sharing and collaborative innovation within the industry, and support the development of a digital infrastructure as well as the enhancement of competitiveness in the entire new materials sector.