Modern information systems based on artificial intelligence (AI) using ML technologies or SIEM systems using AI occupy a significant place in ensuring information security. However, their implementation is accompanied by new threats associated with targeted attacks on machine learning algorithms. The faster AI technologies develop, the faster new classes of attacks on these systems appear. Vulnerabilities of intelligent systems, as well as methods for increasing their resilience are proposed. Particular attention is paid to the development of integrated solutions that combine modern technologies and organizational measures [1]. The paper presents technical and organizational measures to ensure the security of AI systems at all stages of the life cycle, including design, training, deployment, and operation. A structural diagram of a multi-level strategy for ensuring the security of the life cycle of AI systems is proposed, taking into account both cyber threats and specific risks associated with data processing.