Author

Li Jinhua, Doctoral Candidate,Renmin University of China Law School.

Abstract

The inherent characteristics of artificial intelligence (AI), such as its complexity, diversity, and scenario-specificity, have rendered studies on AI tort liability fragmented and lagging behind developments, therefore, there is a pressing need to focus on generalized and universal rules of liability. Compared with AI itself, its developers and users, the providers of AI are the key actors in ensuring prevention and fair distribution of risks. At present, the existing framework of tort law suffers from issues such as internal logical inconsistencies and value disconnection; when applied to cases of infringement involving AI, this may give rise to risks such as unfair allocation of liability, an imbalance between technological innovation and the protection of rights, and blurred internal legal boundaries. The risk characteristics of AI align closely with the risk attribution theory, and risk liability distribution should consider two key factors: the severity of damage and the strength of risk control. Consequently, it is imperative to construct a tiered risk attribution framework that operates in parallel with the fault-based liability system, so as to handle various AI-related infringement cases in a refined, dynamic, systematic, and forward-looking manner, ultimately achieving a fair and reasonable balance of interests.

Keywords:the providers of artificial intelligence;fault liability;product liability;risk-based imputation;safety innovation