AI in Agriculture: Measured Results and Canadian Adoption

Camera-guided spraying cut herbicide 35 to 59 percent in two trials. What Statistics Canada says about drones, GPS, and variable-rate tools.

Guide: AI and SocietyLast verified: October 20264 min read

In farming, artificial intelligence mostly means software that interprets camera and sensor data, and its best-documented use is camera-guided spraying that treats weeds instead of whole fields. A three-year University of Arkansas soybean trial found herbicide use fell 43 to 59 percent compared with broadcast spraying, and an Australian sugarcane trial found a 35 percent average cut. Adoption of the underlying precision tools in Canada is uneven: Statistics Canada reported drones on 3.6 percent of farms in 2020 and GPS on 84 percent of field crop farms in 2021.

Key takeaways

  • Targeted spraying uses cameras and machine learning to find weeds, so herbicide goes only where it is needed.
  • A three-year Arkansas soybean trial reported 43 to 59 percent less herbicide, but a low sensitivity setting let weeds go to seed.
  • In Canada, 13.2 percent of farms used GIS mapping and 16.1 percent used variable-rate input application in 2020.
  • Payback depends on acreage: the Arkansas analysis put it at 819 acres (about 331 hectares) for a USD 25,000 upgrade.

What "AI" means on a farm

Most farm AI is pattern recognition. A camera on a sprayer boom, a drone, or a satellite produces images; a trained model labels what it sees, such as crop, weed, bare soil, or a diseased leaf; and a controller acts on the label. Other uses, like yield prediction or irrigation scheduling, rely on the same idea applied to weather, soil, and machinery data. The software depends on the hardware around it: GPS guidance, sensors, and equipment that can change the rate of a product on the go.

How much of this reaches Canadian farms

Statistics Canada tracks the building blocks. In its 2021 Census of Agriculture article on technology, which reports on calendar year 2020, drones were used on 3.6 percent of farms, auto-steer on 26.8 percent, GIS mapping on 13.2 percent, and variable-rate input application on 16.1 percent. Saskatchewan led for drones at 5.2 percent and for auto-steer at 47.6 percent. A separate 2021 Farm Management Survey found GPS on 84 percent of field crop farms nationally, and among those farms 92 percent used auto-steer.

These figures describe technology, not AI specifically, and the census does not measure machine learning use. They show that the sensors and guidance systems that AI tools depend on are common on grain farms and far less common elsewhere, for example on forage farms, where 32 percent used GPS.

Two measured trials of camera-guided spraying

The University of Arkansas System Division of Agriculture released results on March 10, 2025 from a trial led by weed scientist Jason Norsworthy at Keiser, Arkansas. The release describes soybean plots over three years from 2022. A research team at James Cook University and partners published a separate sugarcane study in the journal Computers and Electronics in Agriculture in 2025.

ItemArkansas soybeansAustralian sugarcane
Herbicide reduction43 to 59 percent versus broadcast35 percent on average, up to 65 percent in low-weed strips
Weed controlHigh setting generally comparable to broadcast97 percent as effective as broadcast
ScaleResearch plots, three seasons25 hectares of field trials
Other findingLow setting raised weed seed bankRunoff herbicide load fell 54 percent

Limits that matter before buying

Savings are not automatic. In Arkansas, the lowest sensitivity setting let larger Palmer amaranth plants survive and seed, raising that population by 280 percent a year, and Norsworthy said the high-sensitivity setting was best in most scenarios. At the high setting, savings are smaller but weed control holds. The release estimated that a USD 25,000 upgrade pays back after 819 acres (about 331 hectares) in a season, and a USD 75,000 upgrade after 2,460 acres (about 996 hectares). Smaller farms may not reach those thresholds.

Results also come from specific crops, weed species, and machines. A trial in soybeans does not tell a Prairie canola grower what to expect, and the sugarcane study measured a custom research sprayer rather than a commercial product.

Checklist for judging a farm AI claim

  • Was the result measured in a field trial, and for which crop and region?
  • Is the saving compared against standard practice or against an untreated plot?
  • Does the claim report weed control or yield, not only input savings?
  • What acreage is needed for the equipment to pay for itself?
  • Was the study independent of the manufacturer?

FAQ

How is AI used in agriculture?

The most documented use is image recognition: cameras identify weeds or crop stress and a controller adjusts spraying or flags a field for inspection. Other uses apply models to weather, soil, and yield data.

How many Canadian farms use drones?

Statistics Canada reported 3.6 percent of farms for 2020, with Saskatchewan at 5.2 percent and Manitoba at 4.9 percent. The agency defines these as unmanned aerial vehicles used for field scouting and data collection.

Does targeted spraying reduce herbicide use?

In the trials cited here it did: 43 to 59 percent in Arkansas soybeans and 35 percent on average in Australian sugarcane. Results varied with weed density, settings, and crop.

Is AI in farming only for large farms?

Not by definition, but equipment costs favor larger acreages. The Arkansas analysis tied payback to hundreds of acres treated per season.

Last verified: October 2026

Sources

  • Canada's farms integrate renewable energy production and technologies, Canadian Agriculture at a Glance, Statistics Canada, May 17, 2023
  • The Prairies and Quebec lead the way in high-tech farm management, Statistics Canada, April 11, 2023
  • Precision agriculture research measures effectiveness of See & Spray technology, University of Arkansas System Division of Agriculture, March 10, 2025
  • Rahimi Azghadi et al., Precision robotic spot-spraying: reducing herbicide use and enhancing environmental outcomes in sugarcane, Computers and Electronics in Agriculture, vol. 235, 2025

New to artificial intelligence?

Start with the basics, then follow the guides to tools, rules, and effects on society.

Start with AI Basics