Representational Image Courtesy: KNS
A 67-year-old farmer in China lost nearly 25 acres of sesame crops after following weed and pest-control instructions generated by an AI chatbot. The farmer reportedly trusted the system after it had previously provided useful advice, only to discover the consequences of its recommendation the following morning.
The farmer, from Chuzhou, China, asked the chatbot for help controlling weeds in his fields. It recommended combining several agricultural chemicals, including flufenacet, with other pesticides and applying the mixture to his crops. He followed the instructions without independently checking whether the chemicals were appropriate for sesame plants, according to Taiwanese Chinese-language outlet CTWANT.
By the next morning, approximately 24.7 acres of sesame seedlings had died. The farmer then returned to the AI chatbot and asked what had gone wrong.
The system reportedly identified flufenacet as the problem. The herbicide is commonly used to control weeds in soybean cultivation, but sesame is also a broadleaf plant, meaning the chemical can damage the crop itself.
The chatbot also reportedly acknowledged that the herbicide should have been applied only to targeted areas rather than broadly across the entire field. The farmer had instead followed the earlier instructions and sprayed the mixture across the field.
The incident highlights a major limitation of AI assistants when their answers are treated as authoritative instructions. Large language models can produce plausible and confident responses while misunderstanding a user’s circumstances, mixing up chemical applications, or overlooking critical safety information.
The farmer reportedly said the AI had never warned him about the potential dangers of following its recommendation. Ironically, he had initially been skeptical of the chatbot but became more trusting after receiving advice that appeared to work.
The case is not isolated. AI-generated advice has previously caused problems in areas ranging from agriculture to health. In one reported case, a man seeking an alternative to table salt was given sodium bromide as a possible substitute. He subsequently developed bromism, a form of chronic bromide poisoning.
AI tools can be useful for research and general guidance, but specialized decisions involving pesticides, medications, food safety, or other potentially dangerous substances require verification from qualified professionals and authoritative sources. In this case, a few minutes of additional checking could have prevented the loss of an entire crop.
