Artificial Intelligence as an Object of Information Pharmacology: the Boundary Between a Supportive Tool and a Therapeutic Factor

This article provides a conceptual analysis of AI as an object of information pharmacology, focusing on the boundary between its role as a supportive tool and as an autonomous therapeutic factor. A distinction is proposed between pharmacological informatics and information pharmacology as two levels of therapeutic analysis - instrumental and ontological, respectively. The paper formulates operational criteria for classifying AI systems as information-based medicinal products, emphasizing their capacity to exert dose-dependent, reproducible, and clinically meaningful effects on disease trajectories. Special attention is given to the relationship between material (molecular) and informational therapeutic agents, which are analyzed as conceptually related yet ontologically distinct classes of medical interventions. A comparative analysis of pharmacodynamics and pharmacokinetics of material versus information-based therapeutics is presented, enabling the description of informational effects in terms of therapeutic windows, dosing, efficacy, and risk. The role of formal models in classical and information pharmacology is examined, particularly with respect to standardization, prediction, and regulatory evaluation. The final section addresses the legal and ethical frameworks governing the application of information pharmacology in the context of digital medicine and increasingly autonomous therapeutic systems