As of 2026, about one farmer in six uses generative AI for farm work, but about one in seventeen names it as a trusted source for decisions. About half of a panel of larger U.S. producers say AI and data-driven tools give their operation no meaningful benefit. Precision agriculture tools, which produce the field records farm-specific AI reads, are mostly on large farms. And the first rigorous trial of AI farm advice will not finish until 2027. Here is where each number comes from, with its date, because these change every year.
01Farmers are trying AI, mostly chatbots.
McKinsey's Global Farmer Insights survey, published in September 2026, asked about 5,500 farmers in ten countries during the spring and summer of 2026. It found 17% use generative AI for farm-related tasks, about 23% in North America. Latin America was highest at 26%; Europe was 13% and Asia 7%. Only about 4% pay for a tool. The same survey found robotics, electric machinery, and sustainability software still barely present on farms.
We could not open McKinsey's own report, so these figures come from Bloomberg's and RFD-TV's coverage of it, which agree with each other.
02Few name it as a trusted source.
Asked which sources they trust for decisions and advice, only 6% of the farmers in the McKinsey survey named generative AI or AI search. 56% named technical agronomists, up from 54% in 2024.
0352% of larger U.S. producers report no meaningful benefit.
The Purdue University and CME Group Ag Economy Barometer asked 400 U.S. producers in June 2026, all with at least $500,000 a year in production, about the main benefit of "AI or data-driven tools" on their farm. 52% said no meaningful benefit. About 23% pointed to higher production, 14% to labor savings, and 11% to less risk. Most also said the tools' recommendations would be hard to follow at least some of the time: 63% said sometimes, 22% said often.
In August 2026 the barometer asked 402 producers a different question: which of their own skills AI could improve the most. Strategic planning led at 32%, then financial management at 28% and production at 18%.
04Precision agriculture tracks farm size.
USDA's Economic Research Service, using 2023 survey data, found that among large crop farms (gross cash farm income of $1 million or more) 68% used yield monitors, yield maps, or soil maps, 70% used auto-steer, 45% used variable-rate technology, and 12% used drones. Among small farms, only 13% used yield or soil maps. Use rises with farm size at every step.
Across all U.S. farms, USDA's National Agricultural Statistics Service counted 27% using precision agriculture practices in 2023, and 22% in 2025.
Read it this way
Yield maps, soil maps, and guidance logs are the data layer that farm AI reads. Farms without these tools can still keep records by hand or with simple sensors, but a farm-specific AI tool needs some record of the farm to work from.
05One researcher on the barriers.
Charles Conteh, a public-policy professor at Brock University, led a two-year study of agricultural automation and robotics in Ontario. Writing in The Conversation in June 2026, he argued the same lessons apply to AI: many tools are technically sound and on the market, and what holds them back is an information gap (growers do not know what exists or what fits), a mismatch (new tools do not fit existing equipment, data, and routines), and fragmentation (universities, vendors, extension, and growers working apart).
Two of the three are about data and how tools fit together, not about algorithms. That is the problem open standards like oat-ods exist to make cheaper.
06Rigorous evidence is just starting.
The first large randomized trial of an AI farm-advice app is under way in Nakuru County, Kenya. Researchers from IFPRI, the University of Chicago, and the University of Bordeaux randomized 600 villages to Digital Green's Farmer.Chat, a video nudge, both, or neither. The main outcomes are whether farmers use extension advice, what they know and adopt, and their say in decisions; productivity, income, and food security are measured too. The program runs through December 2026 and the trial ends in March 2027 (AEA RCT Registry, AEARCTR-0017035).
Read it this way
As of late 2026, no rigorous trial has yet shown that AI advice improves farm outcomes. It is being tested now. Claims of proven returns from AI advice are ahead of the published evidence so far.
The shortest version
As of 2026: about one farmer in six uses generative AI, and about one in seventeen names it as a trusted source for decisions. About half of a panel of larger U.S. producers report no meaningful benefit, and more of them expect help with planning than with anything else. The data layer is mostly on large farms. The barriers are information, fit, and fragmentation. The first rigorous trial of AI advice finishes in 2027.
Frequently asked questions.
How many farmers use AI in 2026?
McKinsey's 2026 Global Farmer Insights survey of about 5,500 farmers in ten countries found 17% use generative AI for farm-related tasks, about 23% in North America. Only about 4% pay for a tool.
Do farmers trust AI to make decisions?
Not yet. In the same 2026 McKinsey survey, only 6% named generative AI or AI search as a trusted source for decisions and advice, compared with 56% who named technical agronomists.
Are U.S. farmers seeing benefits from AI?
About half are not yet. In the June 2026 Purdue and CME Group Ag Economy Barometer, 52% of 400 producers with at least $500,000 in annual production said AI and data-driven tools give their operation no meaningful benefit. About 23% pointed to higher production and 14% to labor savings.
How many farms use precision agriculture?
It depends heavily on size. In 2023, 68% of U.S. crop farms with $1 million or more in gross cash farm income used yield monitors, yield maps, or soil maps, against 13% of small farms. Across all U.S. farms, USDA counted 27% using precision agriculture in 2023 and 22% in 2025.
Has AI farm advice been proven to raise yields or income?
Not in a rigorous trial yet. The first large randomized trial, of Digital Green's Farmer.Chat across 600 villages in Kenya, ends in March 2027.