Scientific Foundations

What is AI in Agriculture & Forestry?

AI technology in agriculture and forestry addresses complex challenges in environmental systems, resource management, and productivity: AI is also known as machine learning and smart technology.

• Agriculture (Ag): Technology such as drones, sensors, predictive analytics aid in pest detection, soil monitoring, and irrigation optimization
• Forestry: AI supports forest mapping, biodiversity tracking, and wildfire prediction

How it Works:

AI in agriculture and forestry relies on large datasets:

• Weather, soil, satellite and remote sensing imagery, and management records serve as input to generate user-specific recommendations
• Public, private, and local sources provide diverse datasets
• On-farm and field sensors aid in generating real-time and site-specific measurements

Uses:

AI in agriculture and forestry analyzes data to support decision-making:

• Improves efficiency, sustainability, and productivity through predictive modeling and automation
• Uses soil, weather, crop, and forest data to create region-specific, resilient practice recommendations

AI Transformation:

Land management enhances precision, efficiency, and agriculture resilience. In agriculture AI improves yields, forecasts drought, and optimizes irrigation and fertilizer schedules. For example, the Soil Twin tool uses machine learning to create daily soil maps and forecasts. In forestry, AI monitors forest health and predicts wildfires.

Download the printable AI in Agriculture and Forestry fact sheet.

What is AI in Computer Science?

A discipline dedicated to creating hardware and software systems capable of performing tasks that typically require human intelligence. This includes building algorithms that allow machines to learn from data, reason through complex problems, perceive environments, and make autonomous decisions.

Machine Learning:

This is a subfield of AI. It aims to train models to learn patterns from data and make predictions. Examples of this include weather, soil, and crop health data can be used to make smarter farming and environmental decisions

These tools don’t make decisions on their own — they provide insights that help farmers and foresters make informed management choices.

Tools Used:

Machine Learning models rely on large amounts of quality data, such as weather patterns, soil composition, temperature records, and satellite imagery, and software that enable the systems to identify patterns and relationships that humans might miss. Powerful computers and large data centers create and operate these models. Agriculture and forestry data collection requires many instruments and professionals, including sensors, satellites, lab researchers, and weather stations.

Applications:

Machine Learning implementation in agricultural data sets can help farmers make a variety of management decisions. These tools support — rather than replace — human decision-making by providing accurate insights from large datasets. For example, AI is using automated controls in irrigation systems, choosing plant phenotyping and weed species identification, and monitoring crop health which can help improve farmer decision-making.

Download the printable Artificial Intelligence & Computer Science fact sheet.

What is Crop Physiology?

Crop physiology is the study of how crop plants grow, function, and produce yield. It focuses on:

• How crops grow and develop 
• How crops respond to their environment
• The biological processes that determine crop productivity 

Crop Physiology & AI:

AI is transforming how we study crops at scale and strengthening precision agriculture. AI tools can:

• Measure thousands of plots in a field quickly and objectively
• Detect plant disease earlier than visual scouting
• Accelerate traditional time-consuming measurements

Together, crop physiology and AI enable faster, data-driven crop improvement.

Why it Matters:

Crop physiology connects plant biology to real-world agricultural outcomes. It helps scientists and farmers:

• Develop high-yielding, stress tolerant crop varieties
• Improve the efficiency of water, fertilizer, and other input use
• Understand how crops respond to heat, drought, and other stresses 
• Adapt farming systems to new extremes 

Applications:

Advances in crop physiology — especially when combined with AI — can help:

• Predict crop growth & yield
• Optimize farm management decisions
• Support breeding of higher-yield crops
• Develop resilient crop varieties

Download the printable Crop Physiology fact sheet.

What is Digital Agriculture?

Digital Agriculture uses Artificial Intelligence (AI) and information communication technology to maximize farm production, increase efficiency, enhance sustainability, and improve decision-making.

Steps of Digital Agriculture:

1. Collect data
2. Clean and organize data
3. Define key features or patterns
4. Select, build, and train a model
5. Evaluate model performance
6. Apply the model in the field

How it Works:

A model will receive data (text, image, video, sensor, and/or audio data), perform computations, and produce an output based on what the network was trained to predict.

Tools:

• Random Forest & Support Vector Machine (SVM): Used for interpretable classification and regression on small datasets.
• Artificial Neural Networks (ANN): Used for complex pattern recognition, classification, and regression with large datasets.
• Deep Neural Networks (DNN): A type of ANN that consists of multiple layers of interconnected nodes between input and output layers.
• Recurrent Neural Networks (RNN): A type of ANN designed to handle sequential data by maintaining a memory of previous inputs.

Applications:

• Detecting Aphid Infestations: Used to identify areas with high aphid populations. Fields can be sprayed before exceeding the economic threshold.
• Detecting Wheat Fusarium Head Blight Disease: Photos of healthy and diseased wheat heads are collected and used to train an SVM model, which is then able to estimate disease severity.
• John Deere’s See and Spray™ Technology: Uses cameras to detect weeds in crop fields. Only sprays when needed, reducing herbicide waste.

Download the printable AI and Digital Agriculture fact sheet.

What are AI Ethics in Agriculture & Forestry?

Ethical concerns in AI arise not only in how data systems operate, but also in how they are trained and deployed in real-world settings. For example, data centers consume resources including land, energy, and water, which can put pressure on local ecosystems.

AI models are prone to bias and “hallucinations”, meaning that their predictive power can sometimes result in errors or fabricated outputs. To reduce the risk of prediction errors, they should be trained to self-correct. Models can also be manipulated for human self-interest, such as through data fabrication.

How Ethics & AI Intersect:

With the rapid emergence of new and unregulated technologies such as AI, developers and users must consider the ethical implications of their design and use.

The intersection of ethics and AI involves several factors, such as water use by data centers, the degradation of critical thinking skills, and academic dishonesty among AI users. Laws and regulations governing AI are still under development.

Tools & Data:

It is important that training data represent phenomena accurately, as some common ethical issues stem from training data being unclean and from how they are interpreted. If the data are organized, error free, or “clean”, the algorithm is better prepared to make more equitable decisions. Ethical frameworks or guidance are available for developers and users.

Examples:

• ACM Code of Ethics
• IEEE Ethically Aligned Design
• EU Ethics Guidelines for Trustworthy AI

Applications:

When working with critical data such as crop yields or forest health status, bias-aware systems that incorporate diverse training data across soil types and regions are necessary to mitigate bias. For example, extension professionals noted that yield and pest-detection models trained primarily on large-farm data can produce less accurate recommendations for smaller or underrepresented farms.

Download the printable AI Ethics in Agriculture and Forestry fact sheet.

What is Remote Sensing?

Remote Sensing is the science of acquiring information about the Earth’s surface without physical contact. Satellites, aircraft, and drones carry sensors that can capture reflected or emitted energy —  some visible, and sometimes that the human eye cannot see such as infrared or radar. Maps of fields or forests can be created.

Challenges:

Remote sensing technology faces several key challenges, including:

• Atmospheric disturbances and cloud cover that obscure data
• Spatial resolution can impacts image clarity
• Information gaps caused by snowy or barren terrain

Tools:

Convolutional Neural Networks: A deep learning model that recognizes patterns in images and spatial data. 

• Used for image and spatial data analysis 
• Automatically learn spatial hierarchies of features from input data

Spatio-temporal Analysis: Examines how patterns change across space and time using sequences of satellite or aerial images. Captures:

• distinct crop patterns,
• continuity in fields,
• different greening cycles of crops.

U-Net: A Convolutional Neural Networks architecture designed for precise image segmentation and classification 

• Used for land cover mapping and feature detection.
• Achieves high accuracy with fewer training images.
• Ideal when instantaneous results are desired.

Applications:

Land cover mapping is essential for monitoring deforestation, managing urban expansion, tracking forest fires, estimating river widths for flood prediction, and evaluating agricultural health through crop growth stage, condition, and yield forecasting.

Download the printable Remote Sensing Tools in Agriculture & Forestry fact sheet.

Key Principles of Soil Health:

The key to productive and resilient soil lies in applying these four interconnected principles of soil health. Following these principles help regenerate the soil’s natural function and unlock its full potential. Principles:

1. Maximize Continuous Living Roots
2. Minimize Disturbance
3. Maximize Soil Cover
4. Maximize Biodiversity

Why it Matters:

Soils carry out essential regulating ecosystem services and store a large amount of the world’s carbon reserves. Soil is a resource that if managed sustainably can be passed from generation to generation

Improvements through AI:

Drought can be mitigated by building the capacity of soils to store water and by using AI tools and weather information to decide when, where and how much to irrigate. Carbon sequestration can be improved through adoption of cover cropping, reduced tillage, and  diversification of crop rotations.

Measurement and Data Analysis:

Soil health is assessed by its physical, chemical and biological components. Tracking these metrics shows the impact of the changes over time.

• Soil Organic Carbon & Organic Matter: This health indicator drives water retention, nutrient cycling and carbon storage
• Infiltration Rate & Aggregate Stability: Measures how well soil absorbs water and resists erosion
• Soil Respiration: Indicates microbial activity
• pH & Nutrient Tests: Determines soil supply and availability of nutrients for plant growth

Download the printable Principles of Soil Health fact sheet.