Artificial Intelligence-Driven Risk Prioritization in Automated Web Vulnerability Assessment

The article examines the transition from static models of web vulnerability assessment to intelligent, context-dependent risk prioritization enabled by artificial intelligence. The relevance of the study is driven by the rapid growth in the number of vulnerabilities, the increasing complexity of web application architectures, and the mounting overload experienced by security analysts, for whom traditional scales, primarily CVSS, no longer provide an adequate distinction between formally critical and genuinely exploitable threats. The aim of the article is to provide a theoretical substantiation and conceptualization of approaches to improving the efficiency of automated web vulnerability assessment by integrating probabilistic models, machine learning, and contextual analysis. The scientific novelty of the work lies in the interdisciplinary synthesis of DAST methods, NLP, predictive models such as EPSS, multimodal ML ensembles, and the practical case of the VULNWatch platform within a unified analytical ecosystem. The principal conclusions demonstrate that AI-driven prioritization enables more accurate forecasting of exploitation risk, substantially reduces the share of false positives, decreases the burden on SOC teams, and shifts vulnerability management from a reactive mode to a predictive one, provided that it is mandatorily accompanied by explainable AI mechanisms and regulatory oversight. The article will be useful for researchers, cybersecurity specialists, SOC analysts, DevSecOps engineers, and developers of vulnerability management platforms.