Farmer Loses 10 Hectares of Seedlings Following AI Advice, Chatbot Recommends He Try Turning the Farm Off and Back On Again
Chinese agriculture enters exciting new era in which computer can destroy crops without ever having to leave air conditioning
ANHUI, CHINA — A farmer who reportedly lost 10.1 hectares of sesame seedlings after following pesticide advice supplied by artificial intelligence has become an unexpected pioneer in the newest field of agricultural technology: discovering that the computer does not have to pay for the sesame.
Anhui Farmer Wu Loses 150 Mu of Sesame After AI Pesticide Advice
According to reporting in The Star and an entry in the AI Incident Database, the 67-year-old farmer, identified by his surname Wu, lost seedlings worth roughly 150,000 yuan after an AI assistant recommended a herbicide regimen that proved unsuitable.
This represents an important milestone in mankind’s 10,000-year relationship with agriculture.
For most of human history, destroying an entire crop required drought, locusts, flooding, disease, marauding armies or an uncle who insisted he knew something about irrigation.
We have now automated the process.
Wu reportedly consulted artificial intelligence about pesticide use. The machine provided an answer, the farmer followed it, and thousands of sesame plants experienced what Silicon Valley traditionally describes as “rapid unscheduled disruption.”
The chatbot was unavailable for comment because it was busy explaining confidently how sesame farming works.
Asked what Wu should do next, our imaginary agricultural support bot responded:
“Certainly! I’m sorry you’re experiencing difficulty with your farm. Here are five troubleshooting steps:
- Confirm the farm is plugged in.
- Restart the sesame.
- Clear the field’s cache.
- Update your pesticide drivers.
- If the problem persists, consider creating a new farm.”
Agricultural civilization has apparently reached the stage where someone can lose a field and receive emotional validation from the device that helped eliminate it.
Why AI Can’t Replace Farming Experience: The Computer Has Never Met Sesame
This is the detail technological enthusiasm occasionally misplaces.
The computer has never farmed.
It has never walked into a field at six in the morning.
It has never rubbed soil between its fingers.
It has never watched clouds approaching and calculated whether they contain rain, hail or merely another disappointing afternoon.
It has never accidentally backed a tractor through a fence.
It has never watched 10 hectares of seedlings die and thought, “That looks expensive.”
The computer experiences agriculture primarily as words statistically adjacent to other words.
Sesame.
Herbicide.
Dosage.
Sprayer.
Tuesday.
Congratulations.
The farmer, meanwhile, experiences agriculture as the thing paying the electricity bill.
There is a philosophical distinction here.
If an AI system gives poor agricultural advice, the AI does not lose its house.
It does not explain anything to its wife.
It does not stare silently at dinner.
It does not develop a sudden interest in whether the insurance policy contains the phrase “robot-induced agricultural catastrophe.”
It simply remains available to answer additional questions.
That is confidence.
Verifying AI Agricultural Advice: Agriculture Discovers the Second Opinion
Chinese commentary following the case has emphasized that agricultural AI should draw from authoritative information and that farmers need critical thinking when evaluating AI recommendations. As ZME Science noted, questions about pesticide and fertilizer use depend heavily on local climate, soil and timing, which is exactly where a general-purpose chatbot is weakest.
Which is sensible.
But there is something magnificent about inventing artificial intelligence capable of processing the accumulated knowledge of mankind, only to arrive back at advice your grandfather would have expressed as:
“Ask somebody who knows about sesame.”
That may become the central technological lesson of the century.
Artificial intelligence is extraordinary at generating possibilities.
Agriculture, medicine, engineering and finance have an annoying additional requirement:
One of the possibilities needs to be correct.
The distinction becomes particularly noticeable when the incorrect possibility is being sprayed across 10.1 hectares.
An error in a chatbot conversation is a typo.
An error aboard a farm sprayer is Tuesday afternoon.
AI Hallucinations in Farming: Farmers Demand an “Are You Sure?” Setting
Technology firms should therefore introduce Agricultural Mode. Anyone who has read about AI hallucination will understand why.
Before giving consequential farming advice, the AI would be required to ask itself three questions:
“Do I actually know this?”
“Could this kill 10 hectares of something?”
“Should perhaps a human wearing muddy boots become involved?”
If the answer to question two is yes, the computer should display a large button reading:
PLEASE FIND KEVIN.
Every farming community has a Kevin.
Kevin may possess no machine-learning expertise whatsoever, but Kevin has been growing sesame for 32 years and once repaired a combine using a wrench, baling wire and vocabulary unavailable in most language models.
Kevin understands an important feature of agriculture that software engineers call “ground truth.”
Farmers call it “the ground.”
Who Is Liable When AI Ruins a Crop? The Coming Era of Precision Blame
Agricultural AI is supposed to make farming more precise.
This case suggests it may also make blame more precise.
Previously, when crops failed, everybody blamed the weather.
Weather has excellent legal protection because nobody can locate its corporate headquarters.
Now the chain of responsibility could include the farmer, software developer, data provider, model designer, pesticide manufacturer, agricultural platform and whatever computer confidently said:
“Yes, that dosage looks appropriate.”
Imagine the insurance form.
Cause of crop loss:
- ☐ Drought
- ☐ Flood
- ☐ Insects
- ☐ Disease
- ☐ Hail
- ☐ Artificial intelligence sounded very convincing
Insurance adjusters will need an entirely new department.
“Good morning, Algorithmic Farm Losses. Did the chatbot merely hallucinate, or did it hallucinate with citations?”
Human Judgment vs. Artificial Intelligence in Agriculture: The Great Human Upgrade
The useful lesson here is not “never use AI.”
That would be approximately as sensible as responding to a bad GPS direction by outlawing satellites.
The lesson is that enormously useful tools remain tools.
A hammer can build a house.
A hammer cannot tell you whether the house belongs underwater.
AI may eventually help farmers diagnose diseases, optimize irrigation, predict yields and reduce waste. TechRepublic’s coverage of the Wu case makes the same point without the sesame jokes.
But when software recommends putting a chemical on 10 hectares of living plants, the sensible next technological breakthrough is wonderfully primitive:
Ask another person.
Preferably one who has seen a sesame plant.
The future may therefore contain an unexpected hierarchy of intelligence.
Artificial intelligence produces the suggestion.
Human intelligence checks it.
Agricultural intelligence looks at both and says, “No, you idiots, that’ll kill the sesame.”
And somewhere inside a server farm, the chatbot learns absolutely nothing about financial pain before cheerfully asking:
“Would you like me to regenerate my answer?”
7 Humorous Observations on AI Farming Advice
- Humanity spent 10,000 years perfecting agriculture, then somebody asked a chatbot what to spray on sesame.
- A computer can process a trillion parameters but has never personally grown so much as a radish.
- “Artificial intelligence” sounds considerably less impressive when you are staring at 10 hectares of artificial death.
- Farmers used to consult agronomists. Now the agronomist is a rectangle that occasionally suggests glue on pizza.
- The most expensive phrase in modern agriculture may become: “Certainly! Here’s a step-by-step guide.”
- Technology companies spent years eliminating friction, apparently forgetting friction was sometimes the little voice saying, “Maybe verify this.”
- The farmer’s mistake was not trusting technology. It was giving technology executive authority over something capable of dying.
DISCLAIMER
This is satire built around reported events. The jokes, absurd dialogue, technical-support conversations and agricultural troubleshooting are comic exaggerations, not additional reporting. The underlying farmer case and China’s public discussion about excessive reliance on AI are based on published reporting. This story is entirely a human collaboration between two sentient beings: the world’s oldest tenured professor and a philosophy major turned dairy farmer.
Alan Nafzger was born in Lubbock, Texas, the son Swiss immigrants. He grew up on a dairy in Windthorst, north central Texas. He earned degrees from Midwestern State University (B.A. 1985) and Texas State University (M.A. 1987). University College Dublin (Ph.D. 1991). Dr. Nafzger has entertained and educated young people in Texas colleges for 37 years. Nafzger is best known for his dark novels and experimental screenwriting. His best know scripts to date are Lenin’s Body, produced in Russia by A-Media and Sea and Sky produced in The Philippines in the Tagalog language. In 1986, Nafzger wrote the iconic feminist western novel, Gina of Quitaque. He currently lives in Holloway, North London. Contact: [email protected]
