AI in farming & food
How can a camera tell a weed from a crop at 15 mph?
Farms are becoming some of the most high-tech places on Earth. AI-powered cameras spray only the weeds, apps spot plant diseases from a photo, and self-steering tractors stay within centimeters of their path, all to grow more food with less waste.
What AI is doing right now
Four ways AI is growing our food
Farming has thin margins and big environmental stakes, so small gains in precision add up fast.
Spraying only the weeds
Cameras across a sprayer’s boom scan the ground, AI decides weed or crop, and individual nozzles fire in a split second.
Real exampleJohn Deere’s See & Spray covered more than 5 million acres in 2025, cutting non-residual herbicide use by nearly 50% on average while scanning 2,500 square feet per second. Precision Farming Dealer
Diagnosing sick plants from a photo
Many plant diseases show up as spots, colors or curling on the leaves, patterns AI can learn from labeled photos.
Real exampleThe free PlantVillage Nuru app lets farmers photograph a cassava leaf and get an AI diagnosis on the spot, even without an internet connection. CGIAR
Tractors that drive themselves
GPS and control math already keep tractors on exact paths, so seeds and fertilizer go only where they should. To work with nobody in the cab, a tractor also needs AI to watch for people, animals and obstacles.
Real exampleJohn Deere’s autonomous tractor, revealed in 2022, carries six pairs of cameras. A neural network labels every pixel of their images in about a tenth of a second and stops the tractor if something is in the way. John Deere · Our explainer on tractor steering
Watching crops from space
Satellites photograph farmland every few days. AI turns those images into maps of crop health and estimates of harvests.
Real examplePrograms like NASA Harvest use satellite data and machine learning to monitor crops and support food security around the world. NASA Harvest
Try it yourself
Be the sprayer’s brain
The AI gives every plant a “weed score.” Decide how sure it must be before spraying, and see the trade-off between saving herbicide and missing weeds.
A balance: 13 of 17 weeds hit, 4 crop plants sprayed. Farmers adjust this setting field by field.
✱ weed ♣ crop blue circle = sprayed
The pattern behind every AI tool
Data → pattern → prediction
Every “AI in…” page follows the same five steps. Learn them once and you can understand AI in any field.
Millions of plant photos
Images of crops and weeds at every growth stage, in sun, shade, dust and rain, each labeled by experts.
Leaf shape and color
Weeds and crops differ in leaf shape, color, texture and where they grow in the row.
Spray or not, in milliseconds
On-board computers classify each plant and trigger the right nozzle as the machine drives by.
Farmers set the dial
Farmers choose how sensitive the system is and scout fields to see what it missed.
Less chemical, more food
Lower costs, less herbicide in soil and water, and healthier yields.
Think like a scientist
What AI can’t do here
Knowing the limits is part of understanding the tool.
Look-alike weeds
Some weeds look very similar to the crop, especially when both are small.
Dust, shadows and speed
Bad lighting, dust or fast driving make photos harder to read.
Cost and access
Advanced machines are expensive, which can leave small farms behind.
Local knowledge matters
An AI trained in one region may not know the weeds, soils or crops of another.
Where people stay in charge: farmers and agronomists decide what to plant, set the AI’s sensitivity, scout fields in person, and make the final calls about their land.
Careers
Jobs where people work with this AI
Agriculture needs people who understand both plants and technology.
Agronomist
A crop scientist who advises farmers on soil, seeds, pests and harvests.
Working with AIUses AI maps from drones and satellites to spot problem areas in a field.
Precision agriculture technician
Sets up and maintains GPS, sensors and smart machinery on farms.
Working with AICalibrates camera-guided sprayers and self-steering systems.
Agricultural engineer
Designs farm machines, irrigation and storage systems.
Working with AIBuilds AI into equipment so it can see and react in the field.
Food supply chain analyst
Tracks how food moves from farm to store.
Working with AIUses AI forecasts of harvests and demand to reduce food waste.
Classroom activity
Train a weed spotter
Students photograph real plants on the school grounds, train an image classifier, and find out what makes it fail.
Teacher tip: before class, try moving a few photos from a phone or tablet to the computer that will run Teachable Machine. A shared folder or a cable both work.
- Collect (15 min): In groups, photograph two kinds of plants outside, such as grass and clover. Take at least 20 photos of each for training. Then take 5–10 extra “test” photos of tricky cases (shade, both plants together, a dead leaf, a blurry shot) and keep them separate.
- Train (10 min): Upload the training photos to Google’s free Teachable Machine as two classes and train the model. Leave the test photos out.
- Test (15 min): In the Preview panel, switch the input from “Webcam” to “File” and try your test photos one at a time. Record where the model fails.
- Discuss (10 min): If this were steering a real sprayer, which mistake would cost more: missing a weed or spraying a crop?
Discussion questions
- How many photos would a real farm AI need, and why?
- Who should own the data a farm’s machines collect?
- How could AI help small farms, not just big ones?
Fresh from the lab
Latest research in farming & food
New studies are added every week and explained in plain language. Not every study uses AI, so ask yourself where AI could help. See all research →
New stories in this area are on the way. Browse all research →
Sources for this page
Put curiosity to work
Read it. Try it. Question it.
Explore new AI research, or take a paper-first investigation into your classroom.