Статья посвящена проблемам информированного согласия при сборе нейрофизиологических данных с применением технологий искусственного интеллекта. Определено, что традиционные формы согласия не учитывают рисков для ментальной приватности и когнитивной свободы личности, что требует пересмотра их содержания и формы предоставления. Предлагаются направления дальнейших исследований и практические решения: закрепление запрета на вторичное использование данных, внедрение динамического согласия с правом отзыва, а также переход к цифровым платформам и интерактивным формам, обеспечивающим осознанность и добровольность согласия.

The article discusses the ethical and legal aspects of prescribing psychopharmacological drugs from the perspective of protecting patients' rights. It presents discussions on the limitations of prescribing psychotropic drugs and the boundaries of the right to mental health. The article also discusses the issue of informed voluntary consent for patients and the ways to ensure their will. The article concludes by proposing the establishment of alternative sources of patient consent in legislation for patients with mental health issues. These sources could include medical consultations, which should be formed based on the need for a comprehensive and interdisciplinary approach. A consultation should generally include specialists with different competencies in the diagnosis, treatment, and rehabilitation of mental disorders, as well as consider the ethical, legal, and social aspects of patient management.

The article examines legal safeguards for maintaining medical secrecy when confidential patient information is transferred for processing by artificial intelligence systems. It shows that the use of AI in healthcare complicates the circulation of such information, expands the range of actors who may gain access to the data, and increases the importance of legal certainty as to the lawful basis for transfer, patient information requirements, and the limits and procedures for controlling subsequent use of the information. The paper identifies key problem areas: the insufficiency of one-time informed consent to medical intervention as a universal basis for AI processing, the limited protective value of de-identification when used as a stand-alone safeguard, and the risk of re-identification. The article outlines the main directions for improving the legal framework by clarifying the grounds for data transfer, standards for patient information, and mechanisms for access control and accountability in AI-based processing.