[The National Data Administration held a symposium on the chain owners of high-quality data sets in the industry: to find out the data requirements for artificial intelligence] November 2nd, the National Data Administration organized a symposium on the chain owners of high-quality data sets in the industry on the morning of October 28 to summarize the progress and results of the chain owners' work, exchange experiences and practices, and research and deploy key tasks for the next step. The meeting emphasized that the construction of high-quality data sets in the industry must adhere to the principles of "determining supply based on demand, promoting construction through use, and improving quality through evaluation" and focus on the following tasks. The first is to find out the data requirements for artificial intelligence. Chain owners should base themselves on their own needs, proactively expand external needs such as upstream and downstream of the industrial chain, adhere to demand orientation, and accurately promote the construction of data sets. The second is to play a leading role. Chain owners should first explore and establish profit distribution and business return mechanisms, sort out industry data maps, promote distributed aggregation of industry data, and guide data sets in this industry to be registered in the national data set management service system. The third is to strengthen application feedback. Relying on actual application scenarios, we establish a data reflow and iterative optimization mechanism, continuously collect user feedback, promote dynamic updates of data sets, and continuously improve the adaptability of data and application scenarios. The fourth is to strengthen industrial chain coordination. Give full play to the "chain gathering" role of chain master units, link upstream and downstream units in the industrial chain, expand the scope of cooperation, and promote the construction of industrial ecology. The fifth is to promote access to facilities. The main units of each chain must actively access the national data infrastructure and rely on the facilities to promote the large-scale circulation and efficient utilization of data sets. Sixth, optimize evaluation services. Promote the transformation of data set quality assessment into a tool-based and order-based model, shorten the assessment cycle, reduce assessment costs, and improve assessment efficiency and delivery capabilities.