Medicine in the Shadow of Algorithms: The Challenges of Trustworthy AI

This article critically examines the implementation of artificial intelligence systems in the healthcare sector, focusing on the Trustworthy AI paradigm. The author argues that the technical efficiency of algorithms is not enough – a trust architecture based on transparency, explainability, and rigorous risk management throughout the product life cycle (TPLC) is essential. The article explores international standards, such as WHO and NIST guidelines, as well as FDA and EU regulations, which define medical software as high-risk systems. A key element is the shift from "black boxes" to models that enable physicians to make informed therapeutic decisions, minimizing the risk of bias and algorithmic errors in real-world clinical environments.