Accompanied by a young farmer named Kennedy Kirui, Sammy Selim strolls through the dense green bush on the slopes of a coffee plantation in the village of Solwot in Kericho, Kenya. They stopped at every corner and entered the farm’s coordinates into a WhatsApp conversation.
This conversation took place with Virtual Agronomist, a tool that uses artificial intelligence to provide fertilizer application advice using chat prompts. The chatbot asks a few more questions before producing a report that says Selim has a yield goal of 7.9 tons and needs to use three types of fertilizer in specific amounts to reach that goal. did.
“My God!” Selim said after receiving the report. He planned to use much more fertilizer than the Virtual Agronomist had recommended. “You could have wasted your money.”
In Kericho and other parts of Kenya, AI-powered tools are becoming increasingly popular among smallholder farmers looking to improve the quality and quantity of their crops.
Farmers are accustomed to suffering large-scale crop losses due to pests, diseases, and lack of technical know-how. They used to rely on advice from agricultural extension agents (experts sent by local governments to provide educational services to farmers), but their numbers have dwindled in recent years due to lack of funding.
Selim started using Virtual Agronomist on his 0.4 hectare (1 acre) farm in 2022 with the help of another farmer who had a smartphone at the time. Following that recommendation, his farm produced 7.3 tons of coffee, its highest yield ever. He is optimistic that the new recommendations will work this time as well. “Technology helps,” he said.
Before adopting Virtual Agronomist, Selim simply applied fertilizers using “general farmer knowledge” without knowing the health of the soil, applying different types of fertilizers at different times of the year. I was there. Farm productivity was low. In one season he was able to produce only 2.3 tons of coffee.
In addition, soil samples were sometimes taken for testing at laboratories far from Solwot, but the results could take months and sometimes not arrive at all.
Farmer Samy Selim (center) and two employees of iSDA, the nonprofit organization that developed the Virtual Agronomist app. Photo: Carlos Mureithi
“The big challenge for farmers is that they don’t know exactly what their soil needs,” said the factory manager of Solwot Coffee Farmers Cooperative, which pulps and dries coffee from local farmers. Flora Maritim says.
The story is similar for farmers trying to identify what pests and diseases are affecting their crops.
Musau Mutisha, from Kwa Mwaura village in Machakos district, said he relied on his knowledge to identify pests and diseases, but it was not always accurate.
On a recent sunny morning on his 0.6-hectare (1.5-acre) farm, he stood next to his corn and used PlantVillage, an AI-powered app for diagnosing pests and diseases, to point his phone’s camera at the tattered leaves. Aimed.
The voice assistant told the user where to make the call, identified the pest as a armyworm, and gave advice on how to get rid of it. “Before, it was all guesswork,” he says. “You’re going to spend more money dealing with things you don’t know.”
Both tools work by training AI models based on images and data. PlantVillage researchers fed the model thousands of images of healthy and diseased crops to learn how to identify pests. Meanwhile, Virtual Agronomist researchers trained a model to predict PH and other soil properties using satellite data from across the continent.
There are 7.5 million smallholder farmers in Kenya. However, the country’s extension worker-to-farmer ratio of 1:1093 is far higher than the 1:400 ratio recommended by the Food and Agriculture Organization.
Enoch Chikaba, director of agricultural delivery systems at the Gates Foundation, which supports iSDA, the nonprofit that created Virtual Agronomist, said farmers need information to be successful. He said technology could help fill the gap created by the lack of extension workers. “We believe in the power of digital,” Chicaba said. “It can really, really mess things up.”
According to a report released in July by the GSM Association, most of the use cases for AI in Kenya, Nigeria and South Africa were in agriculture and food security.
The report says the technology has huge potential to support the continent’s socio-economic growth, but achieving this will require efforts to address digital skills shortages and get more smartphones into people’s hands. states.
Both PlantVillage and Virtual Agronomist use a “lead farmer” model, where farmers with smartphones are trained to use the tools on their own farms as well as neighboring plots. PlantVillage is free to use, as is Virtual Agronomist for all crops except coffee. Advice costs KEN 300 (approximately £1.70).
Despite its promise, some scientists have warned about reliance on AI tools in agriculture. Angeline Wiregui, who studies the use of technology in agriculture in East Africa, says most AI training datasets do not include indigenous knowledge, meaning the information they provide is localized. He said it could exclude successful practices.
“Relying heavily on AI tools to set agricultural practices will result in the erosion of long-held and proven indigenous agricultural practices,” said Wailegi, founder and research director of Athens Research Group. There is a possibility that it will happen.”
Boniface Nzivo uses the FarmShield system to monitor greenhouse conditions on farms he works in Machakos County. Photo: Stephen Mukhongi/The Guardian
But for farmers like Boniface Njibo of Mua village in Machakos district, AI is a game-changer. He uses a system called FarmShield to monitor temperature, humidity and soil moisture and advise when to water his cucumbers. This is something he used to struggle with.
“I don’t waste time thinking about how much water I’m going to use,” he said inside a greenhouse used to grow plants. Plants need a steady supply of water. “It’s amazing technology.”
