Feeding the world’s population is in crisis. The number of people on the edge of famine has jumped to 44 million from 27 million in 2019, the UN's World Food Program said this month.
We need to produce 50% more food in the next ten years to feed the growing population, and the challenge is to do it with less electricity, water and tillable land. We are also running out of farm labor, 50% less than ten years ago. Can AI play a role in addressing this looming crisis?
Here's one scenario for AI and agriculture: a drone intelligently scans the land looking for promising spots for planting. It carries the seeds contained in a shell of fertilizer and plants them. I've heard one company claim that the seedling sprouts into a plant 90% of the time. Such a system can handle up to six drones, at once and do the work of 20 people, and locates viable spots in remote areas.
Once the seedlings are planted, they need water and pesticide. AI systems can use computer vision and only dispense where required, potentially using 50% less water and 90% less pesticide. Soil health sensors could give the farmer 10-12 measurements of moisture, humidity, chemistry, etc., to increase the yield and the cash flow for the farmers.
Vertical farming is also possible. Because the environment is closely controlled, the AI models can constantly adjust the pertinent factors. The yields per “acre” (the vertical farm's footprint, with the help of AI, can produce 10x the yield.
Harvesting - some products could be harvested all at once, others daily based on ripeness. John Deere is at the forefront of bulk harvesting, which I’ll cover below. For the smaller harvest, like picking fruit, the machines are probably too costly to break even at this point. There are quite a few harvester robot models, but their deployment is, at the moment, minimal.
Storage optimization - 99% of all produce goes into some kind of storage before it arrives at the grocery store, restaurant or institution. Almost 45% of all produce is lost or spoiled in the storage and transport. But, an AI system can monitor produce and predict when it will spoil, often weeks in advance, providing the opportunity to get it moving before it does.
We also hear a lot about smart refrigerators, but this has nothing to do with the world food crisis.
But how do wheat and grains fit in?
I see one thing wrong with these proposals. They are producing fruit and produce, not the basic staples needed to feed the world, except for AI-powered devices to ease the labor shortage. Some of the devices do demonstrate the ability to raise yields with less water and far less pesticide.
The world doesn’t need perfectly ripened strawberries; it needs wheat, corn, soybeans. With 175 million acres of arable land in the Midwest, farmers primarily produce wheat, corn, and soybeans. Corn and soybeans are two key products for the Midwest. Last year, corn production was 15.1 billion bushels with a yield of 177.0 bushels per acre, 40% of which was exported.
Wheat is produced on the third most acres in the United States, following corn and soybeans. In 2021, the United States produced 1.65 billion bushels of wheat (about 44 million metric tons) on 37.2 million acres. The most wheat produced in a year in the U.S. was 2.5 billion bushels in 1998 and 2008.
Wheat is the second-largest grain worldwide based on grain acreage and total production volume. The global production volume of wheat came to about over 772 million metric tons in the marketing year of 2020/21. This was an increase of about ten million tons compared to the previous year.
The current crisis
The conflict between Russia and Ukraine is an impending disaster to the worldwide food supply. We are just days away from the wheat planting season in Ukraine, and it’s uncertain if it will happen at all. Farmers will be busy defending the country. Material and equipment generally arrive through Black Sea ports, which at this writing are impassable.
Though the conflict is not in Russia per se, exporting Russia's wheat crop onto the world market is complicated by sanctions and even to a greater extent, logistics traveling adjacent to a war zone. Russia and Ukraine serve as the breadbasket for countries in the Middle East, South Asia and sub-Saharan Africa that depend on imports. Many will be hit hard as a result.
"Any serious disruption of production and exports from these suppliers will no doubt drive up prices further and erode food security for millions of people," the Agricultural Market Information System said in a recent report.
My take - we need AI Solutions for large-scale farming
What should be relevant is a feasible role for AI in agriculture for the primary feedstocks to feed the world. Corn, soybeans, wheat are not panted one delicate seedling at a time. They are planted on hundreds of thousands or more acres at a time and harvested by giant machines that have, until recently, been dumb. The harvesters, "combines" as they are called, require constant adjustment by the driver as they operate, something an AI model would be well-suited for.
John Deere has invested heavily in such programs. Their monster X-seres combine has computer vision to keep the line straight, to adjust height and a dozen other things that required the operator to do manually. Though they are expensive, at over $100,000 apiece, not including all of the other implements, customers say they save at least 40% of their time. Deere is producing a large scale “see and spray” implement, a gigantic autonomous tractor, and a host of other AI-enable products.
These products are not going help the small farmers, but they can boost productivity, with less water and much less pesticide for the large farms (only 29 percent of the global production of crops for food, animal feed and fuel come from land cultivated by smallholders according to Our World in Data).
AI will permeate agriculture in large-scale farming. Its impact on smallholders will remain small, except for fruit and produce farmers.