China's agriculture, vital for food security, faces challenges of pollution and inefficiency due to outdated practices and an aging workforce. While AI is recognized for boosting productivity and reducing emissions in other sectors, its specific impact and mechanisms for driving agricultural green development (AGD) remain underexplored. Previous research has largely overlooked AI's spatial spillover effects and the channels through which it operates.
The study addresses these gaps by examining how AI contributes to AGD in China, focusing on its particular effects across regions and how it operates through two key mechanisms: strengthening human capital and enhancing technological innovation.
Theoretical Framework and Hypotheses
This study established a theoretical framework to analyze how AI drives AGD and proposes four core hypotheses.
The first mechanism operates through the enhancement of human capital. As AI automates routine tasks, it generates a strong substitution effect that increases demand for better-trained workers. This shift pushed labor away from low-skilled, manual roles and toward higher-skilled, knowledge-based occupations, gradually raising the overall level of human capital in rural communities. Such human capital is a critical determinant for improving agricultural technical efficiency and environmental awareness, as producers with higher education are more inclined to adopt green production factors and methods, thereby contributing directly to AGD.
The second channel is to improve “technological innovation capacity.” AI serves as a key driver, making it easier to share and access knowledge, helping agricultural practitioners overcome traditional information barriers, and supporting the development of internal knowledge systems within organizations that can be continually updated and applied. This enhancement in knowledge sharing drives technological innovation, particularly in green and sustainable technologies, which are crucial for reducing emissions and improving production efficiency in agriculture.
Furthermore, the study proposed a non-linear relationship altered by government intervention, introducing the “threshold effect of the level of financial support for agriculture.” While fiscal support can improve production conditions and promote smart machinery, excessive expenditure may lead to over-reliance on government subsidies, potentially distorting factor markets and incentivizing the increased use of chemical pollutants, weakening AI’s positive impact on AGD.
Finally, because technological innovation often functions to serve the public and environmental pollution tends to spread beyond local boundaries, the researchers proposed the existence of a “spatial spillover effect,” where the benefits of AI-driven green practices in one region can extend to and influence neighbouring areas. It proposed that AI’s benefits for AGD are not confined to a local area but can positively impact the agricultural development of neighboring regions through knowledge spillover and policy demonstration effects.
To test these hypotheses, the paper constructed a series of econometric models, including a benchmark regression, a mediating effect model to examine the human capital and innovation channels, a threshold model to assess the role of financial support, and a spatial Durbin model to capture spatial spillovers.
Empirical Analysis of AI's Impact on AGD
This study employed a slack-based measure (SBM) model to accurately assess the level of AGD in China's Yangtze River Economic Belt from 2011 to 2023, accounting for input-output slack and undesirable outputs like carbon emissions. The analysis revealed a consistent upward trend in AGD across most of the 104 cities studied, though significant regional disparities exist.
The upper and middle reaches of the river generally exhibit higher AGD levels than the lower reaches. This is attributed to the more intensive, mechanized, and chemically-dependent farming practices on the plains of the lower reaches, which contrast with the lower energy consumption and emissions in the topographically constrained upper regions.
Empirical results from a double fixed-effects model confirm that AI is a significant positive driver of AGD, a finding that holds under a series of robustness checks. These tests included replacing the core explanatory variable with AI patent counts, employing a dynamic panel model, adding further control variables, and applying winsorization to eliminate outliers.
Mechanism tests demonstrate that AI promotes AGD through two primary channels: enhancing human capital and boosting technological innovation capacity. Furthermore, the threshold model shows that governmental policy creates a non-linear effect: while AI generally promotes AGD, its positive influence weakens once financial support for agriculture passes a certain level, indicating diminishing marginal returns.
When the level of financial support for agriculture exceeds a threshold value of 0.4120, AI's promoting effect weakens, suggesting that excessive fiscal expenditure may lead to inefficiency and reduced innovation incentive.
This research also confirmed that AI significantly boosts AGD locally and in neighboring areas, with spatial spillover effects stronger than direct impacts. The effect is most noticeable in central/western China and major grain-producing regions, highlighting the importance of policy support and favorable resource endowments for maximizing AI's benefits. By contrast, the eastern region shows weaker effects, reflecting its small-scale agricultural operation and stricter environmental regulations.
Conclusion
In conclusion, this study demonstrated that AI significantly promotes (through statistical associations rather than strict causality) AGD in China, directly within local regions and through positive spillover effects on neighboring areas. Its impact is channeled primarily by enhancing human capital and boosting technological innovation.
A critical finding reveals that while financial support for agriculture is beneficial, its positive effect diminishes beyond a specific threshold, indicating that excessive subsidies can be counterproductive. The results show notable regional variation, with the most substantial impacts in major grain-producing zones. These findings support targeted AI integration, strategic fiscal policy, and inter-regional cooperation to advance sustainable agriculture.
Download your PDF copy now!
Journal Reference
Han, S., & Sun, X. (2025). Research on the impact of artificial intelligence applications on agricultural green development. Scientific Reports, 15(1). DOI:10.1038/s41598-025-12836-4
https://www.nature.com/articles/s41598-025-12836-4
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.