Rock typing plays a pivotal role in reservoir characterization by categorizing reservoir rocks into distinct units based on their petrophysical and geological properties. This is essential for accurate reservoir modelling, simulation, and enhanced hydrocarbon recovery. Conventional techniques, including the Hydraulic Flow Units (HFU) and Winland R35 methods, have provided foundational insights over the years but frequently fall short when confronted with the inherent heterogeneity, non-linearity, and data volumes of modern reservoir datasets. The integration of machine learning (ML) has emerged as a transformative paradigm, offering robust data-driven tools for classification, prediction, and spatial extrapolation that complement and extend traditional petrophysical workflows. This comprehensive literature review focuses on methodological advances in ML algorithms applied to rock typing, strategies for integrating these techniques with traditional petrophysical methods, and the emerging challenges that hinder widespread adoption. Key methodological advances include unsupervised clustering for electrofacies delineation, supervised neural networks for permeability and porosity prediction, and deep learning models for image-based rock analysis from micro-CT and thin-section imagery. Integration strategies encompass hybrid workflows combining ML with well logs, seismic attributes, core measurements, and Nuclear Magnetic Resonance data. Emerging techniques, including explainable AI and physics-informed machine learning, are presented. Finally, challenges such as data scarcity, model interpretability, computational demands, wettability representation, and geological generalizability are critically assessed.