Machine learning in ensuring food security: domestic and international practices

Food security remains one of the most critical global challenges, especially in the context of climate change, geopolitical instability, and rising economic inequality. In recent years, machine learning (ML) has emerged as a powerful tool to support data-driven decision-making in agriculture, food systems, and public health. This paper presents a literature review of domestic and international practices in applying machine learning methods to food security, structured around the four key pillars: availability, access, utilization, and stability. The review is based on the analysis of over 35 peer-reviewed publications from 2015 to 2025, retrieved from major databases including Scopus, Web of Science, IEEE Xplore, and AGRIS. For each application area, we examine the most commonly used ML algorithms, data sources, and validation approaches. The results show that tree-based models, deep learning architectures, and hybrid statistical-ML techniques are widely used for tasks such as crop yield forecasting, food price prediction, malnutrition detection, and early warning systems. Despite promising progress, key challenges persist. These include limited data availability in low-resource regions, lack of model interpretability, weak generalizability across contexts, and underrepresentation of Central Asia in global research. The review concludes with practical recommendations for enhancing the use of machine learning in food security monitoring, especially in the context of Kazakhstan and neighboring countries. This study contributes to the growing body of research at the intersection of artificial intelligence and sustainable development and highlights emerging opportunities for context-aware, ethical, and transparent ML applications in food systems.