What AI Down on the Farm Could Mean for Rural America
Brian Carroll lifts his hands, sweeping his palms across a field several green tractors have just begun tilling up, and paints a picture. “Wouldn’t it be great to look like the Epcot Center in Florida, where you have the experimental prototype community of tomorrow?” he asks. “Except here you have the farm.”
Carroll is spearheading an effort to build an agricultural Tomorrowland on 40 acres of black dirt just outside of Horace, North Dakota. The director of operations at Emerging Prairie, a non-profit dedicated to fostering entrepreneurship in North Dakota, he’s overseeing Grand Farm, an initiative launched at the end of April that is in the running—competing with indoor operations like Iron Ox—to create the world’s first fully autonomous farm by 2025.
Farm may be a misnomer, however, because while Grand Farm will include some fields growing crops like corn, soy, and wheat, even those will most likely be for field testing more than actual production. Described as a “a laboratory for entrepreneurs and ag tech companies,” the effort will include a business accelerator, robotics lab, makerspace, and code school called Emerging Digital Academy.
Autonomous farming—marked by equipment fueled by artificial intelligence, using machine learning to gather and compute data to determine and optimize when, where, and how much to plant, fertilize, and the like—has been evolving for years. It has already yielded advancements in precision agriculture and smart equipment such as driverless tractors. Now, projects like Grand Farm promise to take it further.
Grand Farm hopes to advance autonomous ag tech as an approach to addressing several issues facing North American farmers and rural communities today: a skills gap, a lack of venture capital across rural America, and a farm labor shortage pervasive across the U.S. and Canada.
Grand Farm’s primary goal is to provide technology that will take human labor out of the equation, but it’s not an all-or-nothing proposition for Carroll. “I look at it this way: if we get to a fully autonomous farm, that would be great. But to me, the challenges are addressing those pain points… those are more important measurements to see if we’re successful or not,” he says.
The project will develop and test ag-focused innovations, ultimately marrying them with investment opportunities to shepherd them into the wider market. For now, production will center on row crops, but Carroll says the initiative has already gotten interest from mushroom growers and vertical farmers. Grand Farm, which received an investment of 1.5 million from Microsoft earlier this month, also has a bigger goal: to ensure that when the future of agriculture arrives, North Dakota—and the Corn Belt more generally—will have a place at the table.
Like other, similar efforts, Grand Farm also raises important questions about the role of ag tech, and artificial intelligence in particular, at a time when farming has become dominated by large producers and even larger agribusiness interests. Artificial intelligence-based (AI) ag tech also brings with it sociocultural issues that have profound implications for farmers in rural America and the communities they sustain.
Fertile Ground for Innovation
For an experiment that pairs one of America’s oldest industries with one of its newest, North Dakota is fertile ground. Ninety percent of North Dakota’s land is devoted to agriculture; the state is a top producer of row crops including sunflowers, lentils, and soybeans, and 40 percent of North Dakota’s farms weigh in at 1,000 acres or more in size.
Grand Farm isn’t the state’s first foray into farm-tech innovation. The now-international, multi-billion-dollar farm equipment company Bobcat was founded in North Dakota in 1947. More recently, Fargo-based Myriad Mobile launched the app Bushel that digitally handles everything from scale tickets to cash bids for modern grain farmers.
Although Grand Farm’s founders face an uphill battle when it comes to finding investors (more than three-quarters of venture capital went to companies in San Francisco, Los Angeles, New York, and Boston in 2015) the founders hope that an infusion of capital and high-tech jobs could revitalize an area that has been struggling to keep young people from moving away.
AI farming promises a myriad of benefits, such as using precision ag to reduce farming’s carbon footprint through a highly efficient use of resources—everything from fertilizer and pesticides to water and seeds. Autonomous tractors and other machines that use sensors, algorithms, and image-recognition software all have the potential to make farming more efficient by killing weeds, harvesting crops, and spotting crop diseases before they spread.
“I think one of the big benefits from a global perspective is that we’re going to be able to use these technologies to produce more food with less inputs on smaller amounts of land,” says Evan Frasier, director of the Arrell Food Institute and research chair in Global Food Security at Canada’s University of Guelph.
“There is a great interest and desire in technology and where it may go,” says Mark Watne, president of the North Dakota Farmers Union. A recent effort by Google supports Watne’s claim, the company helped convene an event in late 2018 that brought together a number of stakeholders to explore how technology, including AI and machine learning, could accelerate the creation of a more sustainable, scalable, and equitable food system. “We see big opportunities for big data to help small and mid-scale farmers,” according to a report summarizing the event’s findings.
Yet, Watne wonders whether or not the efficiencies and decreased labor costs that AI agtech aims to bring are enough to justify the price of the equipment. “The questions of how fast, what we can afford, and does it actually bring efficiencies, are always the challenge,” Watne says. He hopes that Grand Farm’s research element will be able to explore these cost-benefit issues so that farmers won’t have to take on all the economic risks themselves.
However, there are serious issues facing AI farming that warrant Watne’s cautious stance, from data ownership to national security threats, all existing in a fast-moving tech sphere that far outpaces the ability of lawmakers and regulations to keep up.
Betting the Farm on Big Data
Chances are, the cost concerns that Watne raises are going to be a downstream issue in the data-centered revolution that agriculture is undergoing through initiatives like Grand Farm—a revolution that could make ownership as we know it a thing of the past.
“In the digital agricultural revolution, what we’re talking about is data,” Frasier says. “And in this regard, there’s actually nothing different about farming than any other sector of the economy that’s moving into a data-driven system.”
Consider the last time you upgraded your smartphone. Maybe you walked into a store, forked over some cash, and walked out with a phone you own. But when you powered up the new device and accepted the long list of conditions wrapped up in the phone’s end-user license agreement (EULA), you were consenting to license the software that the phone requires to operate. Through that consent, the company retains the rights to, and control over, the essential components of the phone. What you actually own is merely a chunk of parts that’s largely useless without the software designed to run it.
Many companies in the ag tech sphere are making similar moves to ensure that ownership of the software that powers smart farm equipment stays in the hands of the company, rather than the farmer, through similar EULAs.
“If I’m a farmer and I’ve got AI systems, what’s happening to my data—who does it belong to and what’s it being used for?” asks David Rose, an associate professor in Agricultural Innovation and Extension at the School of Agriculture, Policy and Development at University of Reading in southern England. “Is it being used to maximize benefits on the farm for me or is it being used by a company that wants to sell me more stuff?”
As smart farm equipment churns across a farm, scooping up data on everything from rainfall and water stress to microclimate information and chlorophyll levels along the way, that data can then be used by the company to market products to farmers in the form of AI-fueled recommendations like applying more fertilizer, which said company just happens to manufacture, Rose explains. “If you aggregate data across certain types of farms, you can build up a detailed picture of how that farm works, what products they need, and know that farmer better than they know themselves.”
Data ownership gets particularly murky when data sets are created by a company, corrected by farmers in the field, and perhaps partly owned by the government—should states get involved in the Big Data game. “There’s a huge quagmire,” Frasier explains, one in which farmers aren’t free or able to use the data they generate across different equipment manufacturers and their respective clouds. This results in more ownership and control for companies, and the evaporation of notions of ownership as farmers have known it.
Little has changed in terms of software ownership since the days of the Digital Millennium Copyright Act of 1998, designed to protect movies and music from pirating in the Napster era. “In little bits and pieces, that protection was extended to software that runs devices. Lo and behold, 20 or 30 years later, software is actually on everything. Suddenly this really narrowly tailored legal language… can be applied to tractors on a farm,” says Paul Roberts, the founder of securepairs.org, a group of security professionals advocating for the right to repair. The shift has effectively been turning farmers into tenants rather than owners. And as ownership becomes hazy, so does accountability.
“If a computer [makes a farming decision], what can farmers do if it’s wrong?” Rose asks. “Where does accountability lie? If company says, ‘Apply X amount of nitrogen because it will maximize yields,’ but then a weather event happens that can’t be anticipated that changes the nitrogen necessary and the crop doesn’t turnout how you expected. Whose fault is that and what happens? Does the company pick up the bill?”
The Price of Progress
Beyond farms themselves, the tech-fueled farming revolution is poised to have staggering ripple effects on the rural communities that surround and rely on them as smaller, local farms struggle to keep up. The North Carolina hog industry offers a look at the ongoing relationship between agriculture and technology says Wyatt Fraas, assistant farm and community director at the Center for Rural Affairs.
As technology evolved to allow for the confinement of larger numbers of hogs, the investment required to participate became largely an option only for big farms with the necessary capital. Because one large shipment of pigs is easier to work with than several smaller loads from multiple farms, the market shifted in favor of large farms. “The processors only wanted to work with the larger operations—gave them preferential prices and eventually [processors] became owners of the animals and created the contract systems that we have now where they control all aspects of production,” Fraas says.
As the larger hog farms prospered and offered lower prices, most smaller, local operations were squeezed out of the market. Between 1992 and 2009, there has since been a 70 percent decrease in hog farms across the country, while the number of animals produced has remained roughly the same. This shift has also led to a system in which hog farmers have largely morphed from businesspeople in charge of their operations to little more than landowners who supply companies with a place to mature their pigs.
When technological advancements that come in the form of proprietary software on every conceivable farm device are combined with weak antitrust enforcement and oversight, particularly regarding tech acquisitions, it can lead to consolidation. It’s a quick slide from there into monopolies that “can do what any monopoly does, which is increase costs and decrease service and availability—so you pay more for crappier service,” Roberts notes.
Most American commodity farmers already understand this; they’ve already lived through the advent of consolidation in the seed industry.
“You can grow seed from your own bins as long as you’re not doing something with the genetic trait that’s protected,” Watne says. “But all the new technology out there is being held pretty much in three companies, and we’re paying more for our seed than we ever have.”
As technological advancements that push the farming-as-a-service model proliferate across the farming industry as a whole, “the larger [farms] are more and more going to be leasing equipment rather than owning it,” Fraas says, eliminating the need for local businesses that traditionally serve local farms while enabling the transfer of ownership, and the power that comes with it, to companies rather than people.
Larger farms have less economic activity in the community, fewer farm laborers and owners, and “some less numeric qualities such as less feelings about ownership in the community or interest in participation in the community,” Fraas says of research on industrial farming in rural communities. The opposite was found to be true in communities still buoyed by small- and medium-sized farms.
“What that does is in your local community you end up with fewer farmers and ranchers, and you have fewer people coming to town to do business,” Watne explains. “As those farms grow, they tend to buy further away which, again, hurts that local business. Then, you see these small towns starting to dry up,” unable to sustain local businesses, access to healthcare, grocers, and more.
“It’s serious, and it begs the question if that is what we really want to happen,” Frass says. That ag tech and efforts like Grand Farm will play a central role in this determination is undeniable.
‘Is That the World We Want to Live In?’
“Grand Farm sounds absolutely fantastic,” Rose says, but in the next breath he cautions that it’s important to ask the deeper questions that can stem from this kind of innovation. “It’s the Wild West, when you don’t have legal or ethical guidelines to govern [technological advancement].”
Both Roberts and Rose have taken the issues surrounding software and data ownership and more that recent technology has raised as an opportunity to interrogate the kind of world we want to live in and to start advocating for that vision of the future now.
The right to repair movement that Roberts is involved in—which, at its core, challenges the notions of eternal corporate ownership that software has created—is sweeping across industries, agriculture included. Their effort to fight back against unfair terms that limit farmer’s ability to repair, and thus effectively interact with their equipment, is a noble one, but their path isn’t easy, as manufacturers out-lobby equipment owners 28-1. While corporations have resources—to the tune of billions of dollars spent each year—to pour into lobbying Congress and other federal agencies on their behalf, local farmers simply don’t often have the same ability to get their voices heard.
What the right to repair movement has already accomplished, though, is a glimpse into the regulations that may be required if AI and ag tech are going to be as beneficial to users as they are to the companies that make them. “We need new laws to really make clear that owners own things. Even if they’ve got software on them,” Roberts says. “Anything that the company would use to repair it, the owner should have that, too.”
“I often hate top-down [rules], but when you’re dealing with some of these big players, it’s really only governments that have teeth to do something about it,” Rose adds. Any change, according to Rose, results in winners and losers. For him, this doesn’t mean we should curtail advancements like Grand Farm. But it does mean that extra thought should be given to those on the losing end and put plans in place to support them along the way.
“The future of farming with tech might make more food, but is it all worth it? And is that the world we want to live in?” he asks.