Explainable artificial intelligence as a conditionfor the massive introduction of algorithms in critical areas

Source abstract

This article analyzes explainable artificial intelligence (XAI) as a necessary condition for the widespread implementation of artificial intelligence algorithms in critical areas where the cost of error is high and verifiability of decisions is required. The aim of the study is to demonstrate that explainability is transforming from an optional technical feature into a comprehensive technical and legal imperative that determines user trust, the willingness of specialists to apply AI recommendations, and the legal admissibility of algorithmic decisions. The methodology includes a comparative and analytical review of scientific sources and regulatory initiatives at the international and national levels, as well as a typology of XAI methods. Interpretable models, post-hoc explanations, and hybrid architectures that combine accuracy and transparency are considered. The scientific novelty lies in the justification of XAI as a point of intersection between methods and legal requirements.