Kikis et al. (2026) Benefits and Challenges of Artificial Intelligence in Soil Science—A Review

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Land
• Year: 2026
• Date: 2026-02-15
• Authors: Christos Kikis, Vasileios Antoniadis
• DOI: 10.3390/land15020331

Research Groups

This is a review paper; no specific research groups or departments are identifiable from the provided text.

Short Summary

This review synthesizes recent advances in Artificial Intelligence (AI) applications across key soil science domains, evaluating the performance of various AI methods and identifying critical limitations to their widespread adoption for sustainable soil management.

Objective

• To synthesize recent advances of AI and highlight its applications in key soil science domains.
• To evaluate the performance of different AI methods in soil science.
• To identify major limitations and challenges for the application of AI in soil science.

Study Configuration

• Spatial Scale: Variable, depending on the reviewed studies, covering diverse soil science applications from local to regional scales.
• Temporal Scale: Focuses on recent advances in AI applications, covering various temporal scales depending on the reviewed studies and their specific objectives.

Methodology and Data

• Models used: Random forests, neural networks, convolutional neural networks (as examples of AI methods evaluated in the reviewed literature).
• Data sources: Synthesis of existing literature on AI applications in soil science; discusses the use of large, complex, and heterogeneous datasets, including those from satellite, observation, and reanalysis sources in the reviewed studies.

Main Results

• AI tools are effectively applied in key soil science domains such as digital soil mapping, soil fertility management, soil moisture prediction, contamination monitoring, soil carbon assessment, and precision agriculture.
• AI methods, including random forests, neural networks, and convolutional neural networks, often outperform traditional methods in capturing non-linear soil-environment relationships.
• Major limitations to AI adoption in soil science include data scarcity, issues with reproducibility, lack of large standardized datasets, uncertainty quantification, and the "black-box" nature of many models.
• AI holds strong potential to support sustainable soil management, but its real-world impact depends on better data integration, explainability, standardization, and closer collaboration among stakeholders.

Contributions

• Provides a comprehensive synthesis of recent AI advances and their diverse applications within soil science.
• Offers an evaluation of the performance of various AI methods compared to traditional approaches in capturing complex soil-environment interactions.
• Identifies and critically discusses the major limitations and challenges hindering the broader adoption and impact of AI in soil science.
• Highlights the future potential of AI for sustainable soil management, emphasizing key areas for improvement such as data integration, model explainability, standardization, and interdisciplinary collaboration.

Funding

Not specified in the provided text.

Citation

@article{Kikis2026Benefits,
  author = {Kikis, Christos and Antoniadis, Vasileios},
  title = {Benefits and Challenges of Artificial Intelligence in Soil Science—A Review},
  journal = {Land},
  year = {2026},
  doi = {10.3390/land15020331},
  url = {https://doi.org/10.3390/land15020331}
}