Artificial intelligence in medicine

U radu je dan pregled relevantnih tipova algoritama umjetne inteligencije (eng. artificial intelligence, AI) i njihovih gradivnih elemenata uz osvrt na specifičnosti istih u slučajevima obrade medicinskih podataka. U analizi je dan osvrt na budućnost medicinskih specijalizacija uzimajući u razmatranje najrelevantnije globalne studije, ali i primjere iz hrvatske prakse. Poseban naglasak je dan na kvalitetan pristup iščitavanju znanstvene literature u cilju donošenja informiranih odluka o implementaciji AI tehnologija u zdravstvene sustave sa stvarnim pacijentima. Naposlijetku je dan popis prijedloga za ubrzanje sigurnog usvajanja AI tehnologije u kliničkom okruženju u cilju jačanja konkurentnosti hrvatskog AI sustava u okviru Europske unije i svijeta. Identifikacijom 10 trendova je analiziran potencijal AI-potpomognutog tehničkog unapređenja metoda obrade medicinskih podataka, direktnog kliničkog učinka na specijalističku skrb te globalnog učinka na optimizaciju radnih protokola zdravstvenih sustava, ali i potencijal umjetne inteligencije da u općem smislu: 1. smanji globalni teret bolesti – naročito u segmentu skrbi o kroničnim bolesnicima, 2. osnaži lokalne zajednice, pojedince i inovatore – preuzimanjem veće odgovornosti za zdravlje posredstvom pristupačnih, a moćnih alata te 3. promijeni paradigmu zdravstvene skrbi – od rješavanja akutnih stanja prema ranom otkrivanju, prevenciji i blagostanju (u skladu s definicijom zdravlja Svjetske zdravstvene organizacije (1948.) kao „stanja potpunog tjelesnog, duševnog i društvenog blagostanja, a ne samo odsustva bolesti ili iznemoglosti” [1]). Namjera rada, što je vidljivo i iz raznolikosti pokrivenih tema, je holistički pristup području; a ton je edukativan (s namjerom didaktički) te kao takav potencijalna početna točka dubljeg informiranja inženjera i zdravstvenih djelatnika, ali i zainteresirane javnosti na koju će navedene tehnologije u konačnici imati direktan učinak.This thesis presents an overview of relevant types of artificial intelligence (AI) algorithms and their building blocks with reference to particularities of medical data processing. The analysis provides an overview of the future of medical specialties, considering the most relevant global studies, but also examples from Croatian practice. Special emphasis is placed on a rigorous approach to reading scientific literature to make informed decisions about the implementation of AI technology in healthcare systems with real patients. Finally, a list of recommendations for accelerating the safe adoption of AI technology in the clinical environment is given to strengthen the competitiveness of the Croatian AI landscape within the European Union and the world. An analysis of the state-of-the-art applications of AI in medicine was conducted by identifying 10 common trends showcasing the AI-driven technical improvements of medical data processing, the direct clinical impact on specialist care, and the global effect on workflow optimization of healthcare systems, but also the potential of AI to, generally: 1. ease the global burden of disease – especially in the segment of chronic patient care, 2. empower local communities, individuals, and innovators – by taking greater responsibility for their respective health through accessible and yet powerful tools, and 3. change the healthcare paradigm altogether – from addressing acute conditions to early detection, prevention and well-being (in accordance with the definition of health of the World Health Organization (1948) as a "state of complete physical, mental and social well-being, and not merely the absence of disease or infirmity" [1]). The intention of the thesis, as evident from the variety of topics covered, is a holistic approach to the field; the tone of voice being educational (intentionally didactic) – a potential starting point for deeper dives into the topic by engineers and health professionals, but also the interested public on which these technologies will ultimately have a direct impact