Scale, trust, and the digital divide: a systematic review of AI and ML for agricultural applications
5 affiliations
- 1Data Institute for Societal Challenges, The University of Oklahoma
- 2Department of Geography and Environmental Sustainability, The University of Oklahoma
- 3School of Civil Engineering and Environmental Science, The University of Oklahoma
- 4Laboratory of Geo-information Science and Remote Sensing, Wageningen University & Research
- 5The University of Arizona
Abstract
Artificial Intelligence (AI) and Machine Learning (ML) have transformed the agricultural sector and will continue to do so. Adopting AI/ML technology has the potential to help meet global food, feed, and fiber demands while promoting sustainable farming practices. However, adoption depends on multiple dimensions of trust, particularly trust in system performance and trust in data governance. In this systematic review, we analyze scientific literature for the current status of trust in AI/ML technology for agricultural applications. We use the PRISMA protocol to identify 38 peer-reviewed publications and analyze the content along three dimensions: farming operation scale, trust in AI/ML technology, and the digital divide. Across literature, low trust (66%) and a pronounced digital divide (66%) are dominant themes. Five interconnected clusters, namely data governance, lack of technical skills, lack of explainability, lack of reliability, and high cost of technology, emerge as major barriers shaping trust and adoption. Despite these barriers, farmers across all scales express interest in AI/ML when tools clearly reduce risk or improve yields. However, trust is influenced not only by transparency, but also by structural factors such as data governance, power asymmetries, and unequal access to digital infrastructure. Transparency regarding algorithms' functionality, especially regarding the use of sensitive data, will be important for addressing a lack of trust in the new technology. Training and educating farmers in the use of new tools and technologies, especially at small and medium-scale farming operations, can reduce the digital divide. Targeting intermediate-scale farming operations may further support more inclusive adoption pathways. These improvements could help shift the geographic focus of AI/ML applications in agriculture, supporting progress toward the Zero Hunger Sustainable Development Goal.
Related
- Integrated digital twins for agricultural machinery and robotics: A systematic review of architectures, quantified impacts, and deployment gaps
- Digital divide in information verification among young seniors: An analysis of literacies and sociodemographic factors
- Association between body mass index and musculoskeletal disorders among agricultural workers: a systematic review
- Psychometric properties of the System Usability Scale for digital rehabilitation technologies: a COSMIN systematic review
- Strengthening agricultural extension services to enhance farmers’ capacity and resilience for sustainable development; a systematic review
BibTeX
@article{bhattarai2026scale,
title = {{Scale, trust, and the digital divide: a systematic review of AI and ML for agricultural applications}},
author = {Richa Bhattarai and Jennifer Koch and Wolfgang Jentner and Mehreen Habib and Michael C. Wimberly and David S. Ebert},
journal = {Frontiers in Artificial Intelligence},
year = {2026},
month = jun,
doi = {10.3389/frai.2026.1798896},
url = {https://doi.org/10.3389/frai.2026.1798896},
note = {CC BY},
}
RIS
TY - JOUR
TI - Scale, trust, and the digital divide: a systematic review of AI and ML for agricultural applications
AU - Richa Bhattarai
AU - Jennifer Koch
AU - Wolfgang Jentner
AU - Mehreen Habib
AU - Michael C. Wimberly
AU - David S. Ebert
PY - 2026
DA - 2026/06/19
JO - Frontiers in Artificial Intelligence
DO - 10.3389/frai.2026.1798896
UR - https://doi.org/10.3389/frai.2026.1798896
AB - Artificial Intelligence (AI) and Machine Learning (ML) have transformed the agricultural sector and will continue to do so. Adopting AI/ML technology has the potential to help meet global food, feed, and fiber demands while promoting sustainable farming practices. However, adoption depends on multiple dimensions of trust, particularly trust in system performance and trust in data governance. In this systematic review, we analyze scientific literature for the current status of trust in AI/ML technology for agricultural applications. We use the PRISMA protocol to identify 38 peer-reviewed publications and analyze the content along three dimensions: farming operation scale, trust in AI/ML technology, and the digital divide. Across literature, low trust (66%) and a pronounced digital divide (66%) are dominant themes. Five interconnected clusters, namely data governance, lack of technical skills, lack of explainability, lack of reliability, and high cost of technology, emerge as major barriers shaping trust and adoption. Despite these barriers, farmers across all scales express interest in AI/ML when tools clearly reduce risk or improve yields. However, trust is influenced not only by transparency, but also by structural factors such as data governance, power asymmetries, and unequal access to digital infrastructure. Transparency regarding algorithms' functionality, especially regarding the use of sensitive data, will be important for addressing a lack of trust in the new technology. Training and educating farmers in the use of new tools and technologies, especially at small and medium-scale farming operations, can reduce the digital divide. Targeting intermediate-scale farming operations may further support more inclusive adoption pathways. These improvements could help shift the geographic focus of AI/ML applications in agriculture, supporting progress toward the Zero Hunger Sustainable Development Goal.
SN - 2624-8212
ER -