Across Regions

‘The bots are here – and they’re protecting our crops’

Armed with AI-equipped smartphones, African farmers should be able to detect and deal with potato viruses before they get out of hand. This, according to Jan Kreuze, a member of the science leadership team at the International Potato Center (CIP), headquartered in Lima, Peru.

In a recent article published by Scientific American, Kreuze writes: “Whether it’s our jobs, our privacy or our democracies, artificial intelligence and smart technology often seem to pose nothing but threats. But one “threat” we should welcome from the effects of machine learning and new technology is the disruption of pests and diseases that attack crops, cause food shortages and contribute to famines worldwide.”

Rapid strides in open-source digitization mean we have the tools to equip farmers with intelligent systems that can identify pests and pathogens, allowing them to take action long before any losses occur.

This revolutionary use of technology arms smallholder farmers, who are increasingly gaining digital know-how, with the means to identify quickly and accurately when their crop is under threat, and how best to combat that threat.

Kreuze writes that he’s been working with the International Institute of Tropical Agriculture (IITA) and PlantVillage at Penn State University to develop a smartphone application that gives farmers real-time diagnoses of crop disease, using AI algorithms trained with images of diseased crops.

The PlantVillage app, for example, is being trained to detect diseases in sweet potato, which is a hugely important crop in sub-Saharan Africa. Potato blight alone is estimated to cause annual losses of $5 billion, with potato and sweet potato farmers in developing countries losing up to 60 per cent of their yields to pests.

“Farmers don’t need much technical knowledge or literacy to use the app; they simply point a phone at the infested crop and the app will provide an accurate diagnosis using the talking AI assistant, Nuru,” Kreuze writes.

Read the full article on Scientific American here