From an ontological perspective, artificial intelligence possesses the“materialized”consciousness and personality elements that govern behavior through computer programming and modeling, artificial neural networks, and other technical means. It also has purpose-oriented personality elements based on causal functional connections through self-conscious analysis of big data, which suggests that its personality elements are consistent with those of natural persons. From a social perspective, artificial intelligence, based on its own consciousness, can make limited connections with specific people and things in specific behavioral domains, thus possessing limited social personality elements and special characteristics. Therefore, artificial intelligence, with self-consciousness, can control behavior to achieve specific goals, thus having complete criminal responsibility in specific behavioral fields. Based on the limited social personality elements possessed by artificial intelligence, punishment measures for AI crimes, such as mandatory learning labor, destruction procedures, and reduction of behavioral fields can be created to meet the penalty adaptation requirements of AI crime subjects.