The purpose of this article is to investigate the feasibility of recognizing criminological aspects of artificial intelligence and prosecuting it; the importance of the research issue stems from the growing influence of self-learning programs in human life and their rapid development. It is noted that criminal legislation is not distinguished by the promptness of changes covering the emergence of new relations. There is no corpus delicti associated with the commission of socially harmful activities using neural networks, artificial intelligence, or artificial intelligence itself in the statute. The author analyzes the objective signs of committing crimes and deviances in responses related to artificial intelligence. The work proves that the activities of self-taught programs can pose a public danger and harm public relations, protected by criminal law. Artificial intelligence may potentially fully implement the objective side of some criminal actions, and this list will continue to grow in the future. The article discusses the potential for self-learning programs to engage in unlawful activities. The substance of artificial intelligence's guilt reflects the content of human activity's intellectual and volitional parts. It is suggested that artificial intelligence is not separate from criminology but that it will act in its place if recognized as a person.