Artificial intelligence (AI) in precision agriculture and resource efficiency: a bibliometric and natural language processing analysis

Purpose. The purpose of the study is to map the intellectual structure and resource-economic implications of research on AI-enabled precision agriculture and resource efficiency. The paper focuses on how AI, machine learning, Internet of Things (IoT), remote sensing and decision-support systems are linked to the optimisation of water, land, energy, fertilisers, pesticides, labour and other agricultural resources. Methodology. The study applies bibliometric analysis and natural language processing (NLP) to 598 Scopus-indexed documents retrieved with the keywords “precision agriculture”, “artificial intelligence” and “resource efficiency”. The analysis combines publication dynamics, citation indicators, source and country patterns, normalised keyword frequency, VOSviewer keyword co-occurrence and text-data maps, term frequency-inverse document frequency and non-negative matrix factorisation for topic modelling of titles, abstracts and keywords. Results. The results show a rapid expansion of the field: the number of relevant Scopus records increased from 1 publication in 2018 to 297 in 2025. The dataset received 6251 citations, with a highly skewed distribution, and India, the United States and China were the most visible contributors. VOSviewer keyword analysis identifies six thematic communities: AI, IoT and irrigation-control systems; sustainability, analytics and agri-food systems; precision agriculture and predictive analytics; AI, agriculture and robotics; computer vision and disease detection; and remote sensing, unmanned aerial vehicles (UAVs) and wireless sensor networks. NLP topic modelling confirms six corresponding topics: IoT and real-time monitoring; deep learning and plant disease detection; AI-enabled sustainable agriculture and decision support; remote sensing and UAV-based crop monitoring; food security, climate change and digital transformation; and water, irrigation and resource management. Water efficiency is the most explicit resource-efficiency pathway, while fertiliser, pesticide, land, labour, energy and computational efficiencies are distributed across technical and socio-economic clusters. Scientific novelty. The novelty of the study lies in combining bibliometric network analysis, VOSviewer visualisation and NLP-based topic modelling to interpret AI-enabled precision agriculture as a resource-allocation and agri-environmental management system rather than only as a technological domain. Practical implications. The findings suggest that digital-agriculture policies should address not only hardware and algorithms, but also data infrastructure, interoperability, extension services, farmer training, financing, privacy and digital inclusion. For farm managers, the results identify priority areas for investment, including smart irrigation, soil-moisture sensing, disease detection, UAV and satellite monitoring, edge computing and decision-support systems.