Abstract:  Largescale application of artificial intelligence (AI) has triggered some critical legal issues like the assumption of tort liability, the determination of rights and abilities and the ownership of intellectual property rights. To solve these problems, we must first of all clarify whether artificial intelligence has legal personality and what kind of legal personality it has. As to whether AI has legal personality, current theories have offered various answers including denial, affirmation and limited affirmation. Since absolute answers of yes or no to the legal status of AI do not meet the actual needs, a compromise theory of limited legal personality should be adopted. A study of the historical course of how the connotation of legal personality has gradually expanded and how the proposed AI legal personality came into being is functionally necessary to meet the actual needs. Starting from solving practical problems, we should acknowledge the limited legal personality of AI and establish a doctrine of primary liability that can unveil the mysteries of AI liability so as to realize the positive goal of AI technologies facilitating social development through the adjustment of legal system. 

Key words:  , artificial intelligence, legal personality, AI contracts, legal project