This article examines the legal regulation of artificial intelligence as a means of combating crime in Russia. A significant gap in law enforcement is identified: algorithmic inferences are not recognized as evidence in criminal proceedings, due to the risks of errors in systems used for law enforcement purposes. The fragmentary nature of the regulatory framework in Russia is noted, and a analysis of the regulatory approaches of China and Singapore is provided. Based on the research, a set of measures is proposed: mandatory certification of AI solutions, procedural consolidation of inferences, an accountability mechanism for algorithmic errors.