This contribution explores the opportunities and challenges of applying machine learning in the Earth Sciences, with a focus on igneous petrology and volcanology. It begins by outlining the benefits of machine learning, particularly its potential to automate tasks, enhance modelling strategies, and accelerate knowledge discovery. However, integrating machine learning into scientific research also presents significant challenges. Key concerns include understanding what machine learning models learn, ensuring transparency and reproducibility, and improving model interpretability. These issues become especially critical in high-risk contexts such as volcanic hazard assessment, risk mitigation, and crisis management, where reliance on machine learning outcomes can have serious consequences for human lives. We also provide an epistemological assessment and introduce additional ethical considerations, including the risk of over-reliance on machine learning models and the broader implications of geopolitical development plans, laws, and regulations.