Recently, the Shanghai Municipal Administration for Market Regulation, Shanghai Municipal Commission of Economy and Information Technology, and Shanghai Municipal Commission of Science and Technology jointly issued the "Shanghai Artificial Intelligence Standard System Construction Guidelines (2026-2028)", which outlines a clear roadmap for the standardization development of artificial intelligence in Shanghai in the next three years. By 2028, lead or participate in at least 3 international standards, 10 national standards, 20 industry standards, and a participation rate of over 50% in key areas such as artificial intelligence infrastructure support, key technologies, and industry applications; Create a batch of standardized technology organizations and pilot projects for standardized application, with a key standard application implementation rate of over 80%; Promote the dual empowerment of AI and standardization, and improve the level of standard digitization; Preliminary construction of an AI "standard highland" with international influence.

At present, the iteration of large models, multimodal perception, and autonomous intelligent system technologies continues to accelerate, the boundaries of technology continue to expand, and cross domain integration scenarios continue to emerge. At the same time, problems such as algorithm bias, data security, ethical risks, and insufficient industrial interconnectivity are gradually becoming prominent, and there is an urgent need for a systematic, forward-looking, and implementable set of standard rules to guide and constrain them. Shanghai has gathered a large number of leading artificial intelligence enterprises, research institutes, and innovation platforms, with a complete ecosystem of computing power, algorithms, and data industries. It has the industrial foundation and innovation advantages to lead the exploration of artificial intelligence standardization. The issuance of this guideline is not only a key measure for Shanghai to continuously strengthen its core competitiveness in the artificial intelligence industry, but also an important lever for Shanghai to serve the high-quality development of the national artificial intelligence industry and participate in global artificial intelligence governance.

The guide will build a clear, comprehensive, and dynamically iterative framework for artificial intelligence standards. In the field of basic support, we will focus on the layout of artificial intelligence terminology definition, reference architecture, testing and evaluation, dataset specifications, computing power infrastructure, identification and interoperability basic standards, unify the industry's common language, solve the pain points of incompatibility and interoperability between different technology routes and products from different manufacturers, and build a solid foundation for the coordinated development of the industry. In key technological fields, we focus on cutting-edge directions such as general artificial intelligence big models, computer vision, natural language processing, intelligent robots, reinforcement learning, and federated learning. We develop standards for model training, fine-tuning, deployment, and security evaluation, standardize the technology development process, establish quantifiable and reproducible technology evaluation mechanisms, and guide the industry to break away from homogeneous and low-level competition.

The industry application standards will be based on the unique advantages of Shanghai's urban digital transformation, and deeply integrate with the real economy and urban governance scenarios. Targeting key areas such as intelligent manufacturing, smart healthcare, smart transportation, smart finance, smart education, urban governance, and elderly care services, we will establish scenario based standard specifications. Develop application standards for intelligent production lines and industrial large-scale models to empower the manufacturing industry with AI; Targeting the governance of mega cities, we will improve the relevant standards for government models, urban perception intelligence, and autonomous driving scenarios; Taking into account the needs of people's livelihood, clarifying the norms for the use of intelligent medical AI assisted diagnosis and intelligent health care equipment, while encouraging innovation to be implemented, we will also maintain the safety bottom line.