Farmer Loses Nearly 25 Acres of Sesame After Following AI App’s Chemical Advice

A 67-year-old farmer in Chuzhou, Anhui province, China, watched nearly 25 acres of his sesame crop die overnight after he followed weed-and-pest control advice from an AI app. The incident, which unfolded in July 2026, shows how quickly trust in artificial intelligence can backfire when the stakes involve real chemicals and living plants.

Building Trust Over a Year

The farmer, identified in reports as Wu, started using the unnamed AI chatbot about a year earlier. At first he felt skeptical. Over time the tool gave helpful answers on weather, timing, and general farming questions, so his confidence grew. When weeds and pests began threatening his sesame seedlings, he turned to the same app for a full treatment plan.

The AI produced a detailed recommendation titled something like “Hundred Acres of Sesame Grass Control + Pest Control.” It listed specific herbicides and insecticides and instructed him to spray the mixture across the field. Wu followed the advice exactly. He applied the chemicals—reportedly by drone—across roughly 150 mu of land, about 24.7 acres, without first consulting agricultural technicians or testing a small patch.

Overnight Destruction

The next morning the damage was obvious. The weeds were dead, but so were the sesame seedlings. The young plants had wilted and died even faster than the unwanted vegetation. Wu later said that spraying the mixture meant the seedlings would not survive the following day.

When he asked the AI what went wrong, it pointed to one of the herbicides it had recommended. Agricultural experts who examined the field confirmed the problem: the chemical was designed to kill broadleaf weeds and was unsuitable for sesame, itself a broadleaf crop. Applying it across the entire field instead of limited spots made the outcome inevitable.

A Costly Lesson

Wu estimated his loss at around 150,000 yuan. The episode quickly drew attention online and in local reporting as a cautionary example. Many AI tools include disclaimers that their output can be incorrect and should be verified, yet the convenience of an instant answer can lead users to skip that step—especially after months of positive experiences.

Experts note that chemical recommendations demand local knowledge, precise rates, crop-specific safety checks, and often professional guidance. An AI can summarize information or generate a plausible plan, but it does not replace testing, labels, or an agronomist’s judgment. Wu’s case illustrates how one unchecked suggestion can wipe out a season’s work in a single day.

The story serves as a clear reminder for anyone using AI in high-stakes practical decisions: treat the output as a starting point, not a final instruction, and always double-check before acting.

For more unusual real-world stories, technology mishaps, and surprising true events, subscribe to phactsblog.com. Follow Phacts on Facebook and other social platforms for the latest.

• ShareCopied to clipboard