<![CDATA[Micro, Small, and Medium Enterprises (MSMEs) face persistent financial governance risks due to informal accounting practices, limited digital capabilities, and weak anomaly detection mechanisms in conventional Accounting Information Systems (AIS). This study aimed to map the development, intellectual structure, thematic evolution, and research opportunities at the intersection of Machine Learning (ML), AIS, financial anomaly detection, and MSMEs. A longitudinal descriptive quantitative design was employed using bibliometric analysis and science mapping of 202 Scopus-indexed articles and review papers published between 2018 and 2026. Data were processed using RStudio-bibliometrix, Biblioshiny, and VOSviewer through performance analysis, keyword co-occurrence analysis, thematic mapping, clustering, and thematic evolution analysis. The findings revealed rapid growth in this research field, with an annual growth rate of 60.48% and a publication peak in 2025. Four major thematic clusters were identified: financial forensics, ensemble algorithms, deep learning, and the SME operational context. Random Forest and XGBoost emerged as particularly promising approaches for resource-constrained MSMEs due to their efficiency, predictive accuracy, and interpretability. However, the field remains fragmented, computationally oriented, and limited in terms of validation using authentic MSME datasets. The study concluded that future intelligent AIS development should integrate predictive performance, accounting controls, class-imbalance handling, explainable artificial intelligence (XAI), user acceptance, and broader international collaboration to support transparent, practical, and preventive financial governance in small businesses worldwide.]]>
