Čes. slov. farm. 2026, 75(1):24-29 | DOI: 10.36290/csf.2026.005
Umělá inteligence ve službách farmacie - možnosti využití v praxi
- Mladí lékárníci, z. s., Praha
Úvod: V článku budou představeny principy fungování moderní umělé inteligence (AI), její současné možnosti, omezení a praktické využití ve farmacii. Text přibližuje vývoj AI, vysvětluje klíčové pojmy a seznamuje s nejdůležitějšími modely a jejich schopnostmi.
Cíl: Cílem článku je ukázat, jak může AI podpořit farmaceutickou praxi, kde jsou její limity a jaké legislativní a bezpečnostní požadavky je nutné respektovat při jejím používání.
Výsledky: Moderní AI modely umožňují generovat text, zpracovávat data, analyzovat obrázky a pomáhat v rozhodování. Ve farmacii mohou zrychlit analýzu dat (např. Hlášení LEK-13), usnadnit hledání informací či podpořit dohled nad přípravou IPLP. Ukázky potvrzují vysokou efektivitu modelů (např. GPT-5 při zpracování CSV), ale také odhalují riziko chyb a halucinací. Výkonnost modelů se liší – žebříčky (např. Chatbot Arena) ukazují konkrétní rozdíly mezi GPT, Gemini, Claude a otevřenými modely. Trénink computer vision modelu pro počítání kapslí je proveditelný na běžném PC a je schopný velmi vysoké přesnosti.
Závěr: AI představuje významný nástroj pro farmaceutickou praxi, ale nenahrazuje odbornost farmaceuta. Je nutné rozumět jejím principům, limitům, etickým a legislativním rámcům (AI Act, NIS2) a vždy kontrolovat výstupy. Při správném použití může výrazně zvýšit efektivitu, přesnost i bezpečnost farmaceutických procesů.
Klíčová slova: umělá inteligence, strojové učení, farmaceutická péče, počítačové vidění.
Artificial intelligence in pharmacy services - practical applications
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.
Keywords: artificial intelligence, machine learning, pharmaceutical care, computer vision.
Přijato: 19. únor 2026; Zveřejněno: 13. březen 2026 Zobrazit citaci
Reference
- Tokenizer - OpenAI Platform. [cited 2025 Dec 13]. Available from: https://platform.openai.com/tokenizer.
- Epstein R, et al. Parsing the Turing Test (Philosophical and Methodological Issues in the Quest for the Thinking Computer). Dordrecht: Springer Netherlands; 2008. https://doi.org/10.1007/978-1-4020-6710-5
Přejít k původnímu zdroji... - Artificial Intelligence (AI) Coined at Dartmouth [Internet]. [cited 2025 Dec 13]. Available from: https://home.dartmouth.edu/about/artificial-intelligence-ai-coined-dartmouth
- Russell S, Norvig P. Artificial Intelligence (A Modern Approach, Third edition, Global edition, with Ernest Davis a Douglas Edwards). Harlow: Pearson Education Limited; 2016.
- Vaswani A, et al. Attention Is All You Need. Attention Is All You Need. 31st Conference on Neural Information Processing Systems; 2017
- Walsh S, Timeline Of ChatGPT Updates & Key Events; 2025 [cited 2025 Dec 13]. Available from: https://www.searchenginejournal.com/history-of-chatgpt-timeline/488370/
- Kucharavy A. Overview of Existing LLM Families (Threats, Exposure and Mitigation). Large Language Models in Cybersecurity. Springer Nature Switzerland; 2024.
Přejít k původnímu zdroji... - NÚKIB vydal Varování před některými produkty společnosti DeepSeek. Národní úřad pro kybernetickou a informační bezpečnost; 2025 [cited 2025 Dec 13]. Available from: https://nukib.gov.cz/cs/infoservis/aktuality/2279-nukib-vydal-varovani-pred-nekterymi-produkty-spolecnosti-deepseek.
- Liu Z, et al. am-ELO: A Stable Framework for Arena-based LLM Evaluation; 2025 [cited 2025 Dec 13]. Available from: https://proceedings.mlr.press/v267/liu25ak.html
- Lianmin Z, et al. Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena; 2023 [cited 2025 Dec 13]. Available from: https://doi.org/10.48550/ARXIV.2306.05685
Přejít k původnímu zdroji... - Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence and amending Regulations (EC) No 300/2008, (EU) No 167/2013, (EU) No 168/2013, (EU) 2018/858, (EU) 2018/1139 and (EU) 2019/2144 and Directives 2014/90/EU, (EU) 2016/797 and (EU) 2020/1828 (Artificial Intelligence Act) (Text with EEA relevance), No. 2024/1689; 2024 [cited 2025 Dec 13]. Available from: https://eur-lex.europa.eu/eli/reg/2024/1689/oj/eng
- Shaping Europe's digital future: AI Act; 2024 [cited 2025 Dec 13]. Available from: https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai
- Národní centrum elektronického zdravotnictví: Čato kladené otázky; 2025 [cited 2025 Dec 13]. Available from: https://ncez.mzcr.cz/cs/ai-inovativni-digitalni-technologie-ve-zdravotnictvi/casto-kladene-otazky-faq
- Rydrych M. Doplňující materiály; 2025 [cited 2025 Dec 13]. Available from: https://github.com/Lyman911/AI
- Blog - YOLOv8; 2024 [cited 2026 Feb 7]. Available from: https://yolov8.org/category/blog/
