The paper addresses current security issues of the Internet of Things (IoT), analyzes the most recent and significant violations in this area, proposes a variant of an integrated approach both to protecting the IoT devices themselves and to promising methods organization of interaction using 5G (6G) networks. Particular attention is paid to the process of threat modeling in assessing IoT security, as well as the application of machine learning (ML) principles in order to ensure the protection of the infrastructure of the Internet of things. Practical recommendations are given for organizing measures to improve the security of promising IoT networks.