人工智能践行绿色原则的法律-技术治理路径OACHSSCD

The development,provision and use of AI is accompanied by considerable energy consumption and ecological costs,which are exacerbated by the exponential development of foundation models.The need for AI to implement the green principle is becoming increasingly evident.In order to realize green governance,there is a need not only to clarify how green principles can be better applied to AI,but also to incentivize the development and utilization of energy-saving and emission-reducing technological measures more effectively.Legal provisions such as the Green Principle of the Civil Code and the Principle of Minimum Necessity of the Personal Information Protection Law are applicable to the activities of R&D,provision and use of AI,but lack a systematic elaboration.Deployment measures such as model distillation,compression,and acceleration in the R&D,provisioning,and use activities of AI are supposed to constitute technological measures for practicing the green principle,fulfilling social responsibility,and responding to ESG initiatives.These measures are already commonplace in AI(foundation model)related activities,but legal attention,recognition and incentives remain inadequate.The AI legislation proposal and the Ecological and Environmental Code(Draft),which are being formulated and reviewed,should set more systematic and specific rules with more effective incentives for the practice of the green principle in all specific aspects of AI research,development,provision and use activities,especially for the practice of the green principle in foundation model training,so as to promote the in-depth embedding and internalization of the green principle into the technical and organizational practices.Only in this way,can the most relevant technical facts be more fully discovered about the green principle,and a naturally"incentive compatible"governance approach be designed successfully,which have been widely practiced in the industry.