Sci-fi stories portray computers taking over the world and robots turning against their creators. These popular depictions of artificial intelligence (AI) mask the astounding ways the technology is advancing agricultural production.
The notion machines will replace humans is flat-out wrong, experts say. In fact, machine insights for the agriculture industry are only valuable because technology companies are pairing them with the expertise of agricultural professionals.
“Showing a farmer a satellite image with a map of where their crops are not doing well is not very useful to them, because they know that when they go out and survey their fields,” says Mike Warren, co-founder and CTO at Descartes Labs, a startup that synthesizes satellite and weather data to provide weekly U.S. corn and soybean forecasts. Instead, the value lies in finding ways to provide details that might not be visible, such as where the right application of a chemical can really improve yields.
The primary purpose of AI is to provide decision support and management recommendations based on field-level data. Computers add a layer of infrastructure that people don’t possess: algorithms that can analyze, and find patterns inside of, massive data sets.
Human Element. Although fears of a robot uprising are far-fetched, technology providers should develop user- or business-defined guidelines for using AI, says Robbie Berglund, agriculture and energy sales manager for The Weather Company, a division of IBM.
Team members at IBM refer to AI as “augmented intelligence” because human expertise should inform the final decision.
“An artificial intelligence system is simply offering ideas, ranges of options that could be considered, based on the ability to quickly do millions of bytes of research against sources of data that could never be instantaneously researched in the past,” Berglund explains.
In the future, a farmer could take a picture of a diseased plant and upload it to Watson, IBM’s AI platform. Watson’s algorithm could create a diagnosis, factoring in the plant variety, time of year, weather conditions and which pesticides or fertilizers have been applied.
With that type of structured guidance, an agronomist or grower can make the final decision.
Sharing Economy. As farmers share data on crop performance, machines will develop better algorithms. Synthesized national data will lead to improved management.
“The more of these disparate data sources that we can put together, the better that everyone’s models and analyses will be,” Warren says.
Partnerships with producers, tech professionals and Extension specialists from land-grant universities will aid in the process of data sharing, adds George Kantor, senior systems scientist at the Carnegie Mellon University Robotics Institute.
“They potentially get some personal benefits for themselves,” Kantor says about farmers, “but they also get the broader community benefit of pushing this technology further along.”
In the right hands, AI can be a powerful enabler to improve yields and global food sustainability.
“If we can increase yields with integrity and optimize the operation on the field, leveraging the power of artificial intelligence with accountability, we might be able to meet these challenges,” Berglund says.