Application Areas of Artificial Intelligence in Agriculture: A Critical Review and Analysis

Abstract

Artificial intelligence has moved from a peripheral research interest to a central pillar of modern crop and livestock production. This review synthesises recent peer-reviewed literature on the application of artificial intelligence across the agricultural value chain, covering precision farming, crop disease and pest detection, yield forecasting, weed management, agricultural robotics, precision livestock farming, and Internet of Things-enabled resource management. The review also considers the growing use of explainable artificial intelligence and the bibliometric patterns that characterise this rapidly expanding field. Machine learning and deep learning techniques, particularly convolutional neural networks, have delivered measurable gains in disease classification accuracy, yield estimation, and autonomous field operations, while remaining constrained by data scarcity, poor cross-environmental generalisation, and limited interpretability. Precision livestock farming has extended these gains to animal welfare monitoring and reproductive management, and Internet of Things architectures have provided the sensing backbone that links artificial intelligence models to real-time field conditions. Despite substantial technical progress, adoption remains uneven, particularly among smallholder farmers in low- and middle-income regions, owing to infrastructure gaps, cost, and limited digital literacy. The review concludes that artificial intelligence in agriculture has reached a stage of technical maturity in controlled settings but requires further work on model transparency, data standardisation, and equitable deployment before its benefits can be realised at scale. Future research should prioritise lightweight and edge-deployable models, federated and privacy-preserving learning architectures, and closer integration between agronomic domain knowledge and algorithmic design.