The rapid expansion of networked systems and the Internet of Things (IoT) has intensified the need for intrusion detection systems (IDS) capable of identifying both known and novel cyber-threats. While signature-based IDS remain effective against familiar attacks, machine learning (ML) and deep learning (DL) approaches offer greater adaptability to zero-day exploits. This study presents a large-scale bibliometric investigation of IDS research published between 2015 and 2025, using Scopus-indexed documents. We apply a refined bibliometric mapping tool—integrating multidimensional scaling, co-word analysis, and inverse-transform sampling—to visualise thematic clusters, quantify keyword frequencies, co-occurrences, and temporal dynamics, as well as trace the evolution of AI-driven surveillance techniques. Results reveal a pronounced and accelerating convergence of “machine learning,” “deep learning,” and “intrusion detection system” research, alongside rising interest in IoT-centric security. The most frequent keyword pair, machine learning & intrusion detection system, underscores ML’s centrality to modern IDS design. Temporal analysis shows sustained year-on-year growth for all major keywords. We discuss methodological limitations—including database scope and the exclusion of qualitative impact metrics—and highlight future research directions for adaptive, resilient IDS in increasingly heterogeneous network environments.