Machine learning and artificial intelligence techniques have an ever-increasing presence and impact on a wide variety of research and commercial fields. Disappointed by previous hype cycles, researchers and industrial practitioners may be wary of overpromising and underdelivering techniques. This review aims at equipping researchers and industrial practitioners with structured knowledge on the state of machine learning applications in mineral processing: the supplementary material provides a searchable summary of all techniques reviewed, with fields including nature of case study data (synthetic/laboratory/industrial), level of success, area of application (e.g. milling, flotation, etc), and major problem category (data-based modelling, fault detection and diagnosis, and machine vision). Future directions are proposed, including suggestions on data collection, technique comparison, industrial participation, cost-benefit analyses and the future of mineral engineering training.