Impact of Artificial Intelligence on Cyberattack Detection: A Systematic Review of Recent Studies

Abstract

The increasing sophistication of cyberattacks has exposed the limitations of traditional detection mechanisms and encouraged the use of artificial intelligence to develop more adaptive solutions. This study aims to analyze the current state of scientific knowledge concerning the application of artificial intelligence to cyberattack detection by integrating technological, evaluative, conceptual, collaborative, and thematic perspectives. A systematic literature review complemented by bibliometric analysis was conducted through searches of IEEE Xplore, Scopus, Web of Science, SpringerLink, and the ACM Digital Library. The PRISMA approach, exclusion criteria, and quality assessment were applied, yielding 61 studies for the final synthesis. The results show a marked concentration of Python and Scala as programming environments, while Detection Rate and False Positive Rate predominate as evaluation criteria. Conceptual definitions primarily address technical and functional dimensions, whereas applied approaches remain less developed. Scientific collaboration also displays prominent bridging nodes, and the thematic structure combines specialized areas, such as IoT Security and Intrusion Detection, with research lines that remain insufficiently consolidated. In conclusion, the field is progressing toward greater specialization, although asymmetries persist among technical performance, evaluation, conceptualization, scientific cooperation, and thematic maturity.