Mapping the Evolution of Artificial Intelligence in Agriculture: A Large-Scale BERTopic Analysis of Smart Farming, Automation, and Precision Systems (2020–2025)

Abstract

Advances in artificial intelligence (AI) are revolutionizing agriculture through applications in crop monitoring, precision agriculture, automation, and environmental management. With the rapid development of AI in agriculture, there is an increasing need for extensive evaluation to identify emerging trends and potential future directions. This paper provides a descriptive bibliometric and thematic synthesis of 31,452 publications on the application of artificial intelligence in agriculture in Scopus from 2020 to 2025. The trends and structure of topics in the area are analyzed using BERTopic topic modeling alongside thematic synthesis, temporal trend analysis, centrality-density mapping, and evidence synthesis. Six higher-order themes were identified in the analysis, with crop and production intelligence being the most common. Despite the current progress, the topic of crop-related applications of computer vision remains dominant in the research landscape. However, applications in livestock, socio-technical systems, sustainability, and advanced distributed artificial intelligence fall at the periphery of the landscape. Temporal and structural analyses reveal a concentrated research landscape dominated by production-oriented applications, while interdisciplinary and emerging AI domains remain relatively fragmented. The paper also emphasizes the importance of combining transformer-based topic modeling with thematic and structural analysis in multidisciplinary fields. The results have revealed important insights into the direction AI technology is taking in the agricultural sector, underscoring the need for interoperable, explainable, sustainable, and farmer-centered systems for agricultural development.