Crimes committed on the Internet are characterized by latency and difficulty in their detection and disclosure. Offenders act ahead of time, applying new technologies and all complex schemes, combining them with elements of encryption and thorough disguise. Law enforcement agencies are forced to respond to this using operational-search forces, means and methods, including special operational-technical units, and computer programs being developed. This article proposes to use the capabilities of artificial intelligence and Legel tech technology to monitor the Internet in order to detect and solve crimes. The features of the system, which could, without human intervention, on the principle of virus infection of computers and other technical devices that have access to the Internet, check all their contents and, when detecting marked data, signal to the quick response services about suspected crimes (committed, being committed or prepared), are determined. The article focuses on the fact that the so-called “Internet law enforcement” program should not allow to read or guess the semantic load of phrases, interpret expressions or plots, but only highlight and fix the necessary communication segments. Such information will not violate the constitutional rights of citizens, and could be used to conduct further full-scale operational-search measures. The intended program should have various ways of checking sites, pages on social networks, hosts, disks, and various databases. Its work could be observed in real time, as well as the statistics on the results of inspections for further analysis and reporting could be accumulated. Monitoring of the Internet should take place continuously, with a quick response to the detection of suspicious transmitted or stored data. If necessary, the program should be able to block the transfer of data or the operation of a computer device until the circumstances are clarified and the law enforcement authorities remove the lock.