Analytical approach to assessing the effectiveness of products and data in the field of artificial intelligence based on the adaptive method of determining expected utility

The article examines the problem of a comprehensive assessment of the effectiveness of digital products and data arrays in conditions of high uncertainty and characteristic risks in the implementation of intelligent systems. The relevance of the work is determined by the rapid development of cognitive technologies in the absence of universal analysis methods that take into account both technical characteristics and economic indicators. A critical review of existing assessment approaches has been conducted, and their main limitations related to insufficient consideration of the dynamism of the technological environment have been identified. The main contribution of the research is the development of an adaptive methodology for calculating expected utility, integrating quantitative and qualitative indicators through a system of weighted parameters and probability coefficients. The proposed approach allows not only to evaluate the current performance of digital solutions, but also to predict their potential in a changing environment.