Explainability, Robustness, and Fairness in User-Centric Intelligent Systems: A Systematic Review

The demand for tailored user-centric systems is increasing in the evolving landscape of artificial intelligence (AI). This paper systematically explores the literature on user-centric intelligent systems, focusing on three vital dimensions: explainability, robustness, and fairness. By employing a rigorous systematic literature review and adhering to the PRISMA protocol, this study scrutinizes articles from several esteemed online scientific journals—IEEE Xplore, ScienceDirect (SD), Web of Science (WoS), and Scopus. Crafting a coherent taxonomy from insights gained through meticulous analysis, ensuring a comprehensive review. Categorizing the literature by explainability, robustness, and fairness, the resulting taxonomy aids readers in navigating this intricate domain, and each category is investigated through different approaches. Detailed discussions, enriched with insights from challenges, motivations, and recommendations in prior articles, analyse the literature through identified perspectives. This paper explores datasets, methods, and frameworks researchers employ, providing a holistic view of methodologies in user-centric intelligent systems. Going beyond a standard literature review, this work guides readers and future researchers, addressing challenges and advancing user-centric intelligent systems. This paper serves as a significant resource for understanding the current user-centric intelligent systems landscape. As the field evolves, subsequent works can build upon this foundation, exploring the nuanced dimensions of user-centric intelligent systems.