Recently, Jiang Xiaojuan, a professor at the University of the Chinese Academy of Social Sciences, gave a speech at the academic annual meeting on China's digital economy development and governance, "Artificial Intelligence Development and Governance: What perspectives and standards should economics have? 》 Keynote speech, deeply analyzes the key issues of artificial intelligence development and governance from an economic perspective, and emphasizes the need to build a theoretical framework for technological development and rule design with discipline logic, providing new ideas for solving the dilemma of artificial intelligence governance of "focusing on principles, neglecting academic principles, emphasizing concepts, neglecting practice".

　　From moral principles to practical rationality: reconstructing the legitimacy standard for the development direction of AI. For a long time, discussions on the rules and ethics of the development of artificial intelligence have focused on the level of moral concepts and value advocacy. The 12 principles of security, transparency, and responsibility proposed by the Asiloma Principles of Artificial Intelligence launched in 2017, as well as the six priorities such as inclusion and sustainability emphasized in the statement signed by 58 countries at the 2025 Paris AI Summit, which have gathered social consensus, there are significant problems of insufficient academic support and practical transformation.

　　Jiang Xiaojuan said that these principled expressions lack an analytical academic framework and measurable evaluation indicators, making it difficult to scientifically judge the advantages and disadvantages of technology or products, and cannot systematically weigh the implementation costs and social benefits of governance rules, resulting in slow implementation of relevant concepts. She emphasized that the judgment of the legitimacy of the development direction of artificial intelligence should shift from abstract moral principles to practical rationality in economics, and establish an analytical framework based on resource allocation efficiency and social fairness, so that the "destined" goals and "real" effects of technological development form a verifiable logical closed loop.

　　Considering efficiency and fairness both in dual dimensions, build an economic evaluation system for the social value of AI. Jiang Xiaojuan believes that from the perspective of economics, the core of the social value of artificial intelligence lies in whether it can promote "improving the efficiency of social resource allocation" and "equal sharing of development results". She believes that the impact of artificial intelligence on social fairness can be divided into two categories. One is the continuous impact on existing fairness and balance, including large enterprises strengthening their market dominance with their advantages in data, computing power, and algorithms, and squeezing the living space of small and medium-sized enterprises. The second is the disruptive challenges brought by the native characteristics of technology. For example, the large artificial intelligence model presents the characteristics of "increasing returns to scale" and "capacity emergence". The larger the model, the higher the efficiency of optimization and iteration, and after breaking through the threshold, it produces cognitive and problem-solving capabilities that small models cannot have, forming a "crushing" competitive advantage.

　　Jiang Xiaojuan believes that judging the social value of artificial intelligence must go beyond the single efficiency dimension and establish a multi-dimensional indicator system covering intergenerational fairness, group fairness and opportunity fairness to ensure that the results of technological progress are shared by all members of society.

　　Coordinate strong and weak rules to build a dynamic balanced AI governance framework. Jiang Xiaojuan proposed that artificial intelligence governance needs to learn from the logic of the synergy between "market endogenous rules" and "government regulatory rules" in economics to form a governance system that organically combines "weak rules" and "strong rules". Weak rules are the basic constraints of market game and social coordination, while strong rules are the bottom line guarantees of government supervision and legal constraints. For example, implement negative list control, strengthen transparency requirements, and protect the rights and interests of digital vulnerable groups. She specifically pointed out that the impact of the artificial intelligence technological revolution on employment is different from the historical situation where "new positions are more than old positions". It is necessary to establish a more forward-looking social security and vocational training system to avoid the social imbalance of "winner's benefits" that cannot offset the "loser's losses".

　　"Artificial intelligence governance is not a simple rule-making, but a systematic project involving economic efficiency, social fairness, and technological innovation." Jiang Xiaojuan said that the core value of the economic perspective lies in providing quantifiable, comparable and operational analytical tools to transform governance goals from a "beautiful vision" to a "practical path." Only by adhering to the unity of efficiency and fairness, the coordination between the market and the government, and technological innovation and humanistic care can we lay a sustainable institutional foundation for the development of artificial intelligence and truly achieve the ultimate goal of "technology benefits mankind". (Author Fuxi)

[Editor in charge: Zhu Jiaqi]