Key points are not available for this paper at this time.
INTRODUCTION: Paper presents the principles of modern artificial intelligence (AI), its current capabilities, limitations, and practical applications in pharmacy. The text outlines the development of AI, explains key terms, and introduces the most important models and their capabilities. OBJECTIVE: The objective of the article is to show how AI can support pharmaceutical practice, where its limits lie, and what legislative and safety requirements must be respected when using it. RESULTS: Modern AI models enable text generation, data processing, image analysis, and decision support. In pharmacy, they can speed up data analysis (e.g., LEK-13 reports), facilitate information retrieval, and support the supervision of IPLP preparation. Examples confirm the high efficiency of the models (e.g., GPT-5 in CSV processing), but also reveal the risk of errors and hallucinations. Model performance varies-rankings (e.g., Chatbot Arena) show specific differences between GPT, Gemini, Claude, and open models. Training a computer vision model for capsule counting is feasible on a standard PC and is capable of very high accuracy. CONCLUSION: AI is an important tool for pharmaceutical practice, but it does not replace the expertise of a pharmacist. It is necessary to understand its principles, limitations, ethical and legislative frameworks (AI Act, NIS2) and always check the outputs. When used correctly, it can significantly increase the efficiency, accuracy, and safety of pharmaceutical processes.
Milan Rydrych (Fri,) studied this question.