Limiting factors and ethical challenges in the application of artificial intelligence in pharmacology

Introduction: Artificial intelligence (AI) is increasingly applied in pharmacology, particularly in drug discovery, prediction of pharmacokinetic and pharmacodynamic properties, identification of adverse drug reactions, pharmacovigilance, dose optimization, and clinical decision support. Despite these promising applications, its integration into pharmacological research and clinical practice remains limited by scientific, technical, ethical, and regulatory challenges. Aim: This review aims to summarize the main limiting factors and ethical challenges associated with the application of AI in pharmacology and to outline key requirements for its responsible and clinically meaningful implementation. Materials and Methods: A narrative review of recent scientific literature on AI tools in pharmacological research and practice was performed. The analysis focused on limitations related to data quality, representativeness of training datasets, algorithmic bias, model interpretability, external validation, reproducibility, and performance across clinical populations. Ethical and regulatory aspects, including privacy, informed consent, fairness, accountability, transparency, and governance, were also examined. Results: The reviewed literature indicates that AI has substantial potential to support pharmacological decision-making and accelerate innovation. However, models trained on incomplete, heterogeneous, or biased datasets may produce unreliable outputs and reinforce inequities in healthcare. Limited transparency of complex algorithms reduces clinician trust and complicates the evaluation of safety, responsibility, and clinical applicability. In addition, the lack of harmonized regulatory standards remains a major barrier to broader implementation. Conclusion: AI should be regarded as a supportive tool that complements, rather than replaces, expert pharmacological judgment. Responsible implementation requires high-quality data, robust validation, transparent reporting, interdisciplinary collaboration, continuous monitoring, and ethically grounded regulatory frameworks.