What you’ll learn in this article:
- A new report from Data and Society evaluates the impact of AI and predictive analysis technologies on farmers.
- Beyond concerns about employment loss, researchers are assessing the impacts of AI’s human infrastructure and the hidden human labor that helps build AI.
- Lawmakers in the U.S. House, the Trump administration and national governments have weighed in on concerns about the future of work in the age of AI.
Conversations around future-of-work issues often steer us into the abyss of fewer job opportunities, or on the flipside, toward a fantastical future free of grunt work. But whether or not robots will ever replace all human workers, many of us already work alongside AI all the time. Today, researchers are evaluating the effect of AI and predictive tools on farm work, exposing the hidden human infrastructure necessary to train and maintain automated systems — and asking a lot more than the “Will robots take my job?” question.
A recent Wired article headlined “A.I. Is My Coworker” in print, concludes that “If AI handles the most boring tasks, we humans just might have more free time to be creative.” That, of course, is one of the promises of AI: Humans will advance not only as intelligent machines take on our menial tasks and laborious physical work, but as AI enhances our brain capacity to imagine and invent.
Not so fast, say Alexandra Mateescu and Madeleine Clare Elish. The two anthropology researchers recently analyzed the impact of emerging technologies on workers and their everyday tasks and interactions as part of Data and Society’s AI on the Ground initiative. The biggest takeaway for Mateescu? “The risks and benefits of adopting new technologies are not equally distributed.”
A Thorny Reality for AI-Sourced Crop Cultivation
In practice, while some AI-based tools might make work easier (consider those automated responses in Gmail that Wired’s Lauren Goode wrote about, for example), they often require more or different forms of work. The Data and Society report “AI in Context: The Labor of Integrating New Technologies” observes the effects of AI and data-fueled analytics on agriculture. The researchers evaluated impacts of precision agriculture technologies such as GPS-enabled combine harvesters, and software that helps farmers predict and analyze soil makeup to pinpoint fertilizer and pesticide application.
At the level of farm management decisions, this work can include shifts in ingrained community norms, interpersonal relationships, daily routines, and skill sets. Additionally, there is the labor of reconfiguring physical infrastructure to render farmland amenable to the data collection that makes AI possible. This is not always a smooth process, and the extent to which farmers can benefit from precision agriculture tools often depends on the conditions and resources already available.
To incorporate AI or other tech, farmers might need to redesign a barn to facilitate sensors or change crop layouts to fit new machines. However, they also may find that there is a lot of hidden labor necessary to get data to do what they need it to do. Indeed, over the past decade or so, practically every industry has been disrupted by new analytics technologies and processes that, in order to operate well, require data categorization and standardization, de-duplication and massive data management overhauls to enable integration of business intelligence systems.
“The risks and benefits of adopting new technologies are not equally distributed.”
– Alexandra Mateescu, Data and Society
So, data cleanup is necessary in agriculture as well as in marketing or logistics. One key difference? Staring at a computer all day isn’t exactly what most farmers signed up for. It requires some attitudinal adjustment. “Sitting at a computer doing information work is fundamentally different than making physical action happen in the world,” said Elish. “There’s a mental component about being able to make the shift of how you think about the business of what you do and how you think about what it is to farm.”
This AI interruption in the traditional farm workflow also affects the “locus of authority,” said Elish. While the decision-maker was once a human farm owner or manager, automated prescriptive technologies from agtech giants such as Monsanto can create power and influence from new data overlords.
Human Infrastructure and the Hidden Labor of AI
Ultimately, in the current landscape, AI in the workplace does not necessitate easier work nor does it necessitate massive job loss. We hear all the time about the unmet demand for computer scientists and machine learning experts in the AI industry. But all sorts of AI technologies rely on teams of less-skilled workers to facilitate their creation and even help make them work properly after they’re built.
Data and Society calls this “‘human infrastructure,’ the integral human component of a socio-technical system without which that system cannot properly function.”
Human Infrastructure – the integral human component of a socio-technical system without which that system cannot properly function.
In an adjacent discussion, other researchers have analyzed AI’s hidden workforce. In its 2018 report, AI Now Institute refers to the “hidden human labor” of AI, evaluating the behind-the-scenes task-based gig-work or “click-work” needed to label and categorize data used to train AI systems. They and other researchers also have analyzed the surveillance technologies that oversee workers and monitor performance.
Recognizing all of the labor required to “make AI work” can help us better understand the implications of its development and use. Research in these areas also helps us reexamine the focus on technical talent in narratives describing AI’s creation and recognize that technical skills account for only a portion of a much larger effort.
Microsoft researchers and authors of the forthcoming book “Ghost Work,” Mary Gray and Siddharth Suri, wrote about the “people behind the AI curtain” for Harvard Business Review in early 2017. “We call this ever-moving frontier of AI’s development, the paradox of automation’s last mile: as AI makes progress, it also results in the rapid creation and destruction of temporary labor markets for new types of humans-in-the-loop tasks.”
Governments Weigh In on AI and Work
Though they tend to focus on AI’s threat to employment, governments have addressed the impact of AI on labor. Last month, Congresswoman Brenda Lawrence (left), a Democrat representing Detroit, along with Silicon Valley Democrat Ro Khanna introduced proposed guidelines for ethical AI in a House Resolution. And, for Congresswoman Lawrence, it turned out future-of-work was the impetus for her involvement in the proposal.
“Representing the city of Detroit, which has a workforce participation rate of 53.4 percent and one of the lowest rates of internet connection in the country – I realized that advances in technology are helping the few and not the many,” Lawrence told RedTail’s Kate Kaye for an International Association of Privacy Professionals story. “I introduced HRES 153 so we can help shape the dialogue and development of AI to work for all Americans.”
“I introduced HRES 153 so we can help shape the dialogue and development of AI to work for all Americans.”
– Congresswoman Brenda Lawrence
The Trump administration has also weighed in on work in the age of AI. After unveiling a vague and slim AI strategy document in February, the White House launched a more robust microsite on March 19 that serves as a hub for AI initiatives underway at several agencies. In the future-of-work category, the administration appears to be focused primarily on training people for emerging occupations. However, it does mention research at the National Science Foundation intended to “help us better understand the human-technology partnership and the emerging socio-technological landscape, create new technologies to augment human performance, and foster livelong and pervasive learning with technology.”
Other countries including Finland, France and Germany have established initiatives on AI and its impact on employment, particularly in communications related to AI strategy and ethical guidelines for AI. In France, for instance, a 2018 government report addressed job loss in addition to the impacts of AI on work in the nearer-future.
“In light of what may now be considered an inevitability, in the medium-term we need to be pushing on with discussions on alternative modes for producing and redistributing value,” noted the “AI for Humanity” report. “The priority must be on developing the means for effective complementarity between human tasks and machine tasks.”