EXPLAINABLE ARTIFICIAL INTELLIGENCE AS A SERVICE: ALGORITHMS FOR ASSESSMENT OF CHARACTERISTICS

The research focuses on methods for evaluating Explainable Artificial Intelligence as a Service (XAIaaS). An analysis of existing approaches to evaluating explainable artificial intelligence systems has been conducted. The purpose of the study is to develop an approach and a sequence for assessing the quality of explainable artificial intelligence systems within the XAIaaS model, based on a characteristic-oriented methodology. A multi-level evaluation scheme is proposed, which provides step-by-step analysis, testing, and correction based on the defined key characteristics of the first level of the quality model hierarchy. This enables targeted, structured, and comparative evaluation of individual system properties as well as of the system as a whole. The application of the proposed evaluation methodology is demonstrated through a practical example. Keywords: artificial intelligence, XAI as a service, explainable artificial intelligence evaluation, requirements profiling.