Martindale Center- Policy Briefs on the Future of Work

The Economic and Technical Feasibility of AI Substitution of Harvesting Jobs in the United States During COVID-19 2 resolved. In part, this is due to the fact that each country has its own plans to address the pandemic and the collection of data, including within the labor sphere. In any case, although rising unemployment is of great concern, it is a secondary priority for most, if not all, governments at this time, their first priority being the containment of the pandemic itself and ensuring access to medical relief for those affected by the virus. In that sense, most governmental policies during the harshest months of the pandemic initially have been focused on public health measures, including various forms of quarantine, regulations on business openings, and work- from-home requirements. Such measures have drastically affected workers, many of whom have lost their jobs, while others have had much reduced access to their places. According to the Oxford Review of Economic Policy study, “public health measures and changes in preferences caused by avoidance of infection” have become the primary reasons for effects within the labor sphere (del Rio-Chanona et al., S95). The economic crisis caused by the pandemic has had both supply-side and demand-side effects, which secondarily affectworkers andworkplaces.With regard to the supply-side, many industries, small businesses, or independent workers are incapable of carrying out their labor activities due to government restrictions imposed mainly on “non-essential industries and workers.” This “essentiality” measure has resulted in a widespread growth of isolation, where much of theworkhas beendone fromhome. Aconsiderable percentage of workers are on total stop, since their labor activities were unable to adapt to remote alternatives. On the demand-side, people are not consuming goods or services, activities that might increase their risk of contracting the virus. Because of these two sides, most industries’ operations have been drastically reduced and, in consequence, so too have their work forces been reduced. Given the above background, an important, even if seemingly unusual, question can and should be asked. That is, has COVID-19 affected the use of AI robots in the process of harvesting in the US agriculture industry, and if so, to what extent has it affected the economic and technical feasibility of possible substitution for labor in this area of work? Harvesting is one of the least mechanized and digitized branches of agriculture. Most fruit crop harvesting is done by hand, involving seasonal workers who are likely to be unauthorized immigrants engaging in hard physical/manual labor (Calvin & Martin, 41). The seasonal worker position primarily involvesharvesting fruit crops safelyandefficiently, ensuring productivity, and achieving goals that focus on thoroughness, speed, and quality. Restrictions related to COVID-19 had an obvious impact on the agricultural workforce, especially with regard to the pool of seasonal workers who are typically employed in fruit crop harvesting. This has led to a labor shortage and has jeopardized food security (Bochitis et al., 1). The possibility investigated here is that of turning to robotic agriculture as a solution to speed up harvesting (Di Vaio et al., 4). In the US, agriculture is important not only for foodbut also for jobs andsocial fabricof many communities, so agriculture harvesting robots play an increasingly central role. As the ongoingmigration fromrural to urban areas results in labor shortages in agricultural areas, agricultureharvesting robots canhelp“increaseproductivity, reduce waste, and increase agricultural sustainability” (Tang et al., 2). To judge the feasibility of AI complementing or substituting for labor as it pertains directly to the job of a harvester in the COVID-19 crisis requires a brief review based on four data points: the Remote Labor Index (RLI), the “essentiality,” the risk of exposure to disease and infection, and a measure of probability-of-substitution by AI. As background to this analysis, it is important to recognize the overall trends in this industry that have taken place over the course of many years (andwhatwasexpected inthe futurebypre–COVID-19 future projections for the industry) so as to distinguish them from anyfindings/changes in theCOVID-19era agricultural sector. ReferringtotheOccupational InformationNetwork(O*NET), harvesting is included in the category of “Farmworkers and Laborers, Crop, Nursery, and Greenhouse.” Harvesting is designated as having an “average” employment growth, with a projected increase of 3.8% for the decade 2019–2029, according to the US Bureau of Labor Statistics. Comparing this projection with the percent employment change between 2011 and 2019, which was approximately a 39% increase, indicates that the available additional harvesting jobs will be significantly lower compared to the last decade. With this background in mind, a look at the effects of AI and automation in substituting for harvesters tomeet the demand for agricultural products within the context of COVID-19will make more sense. Taken together the first three data points nicely categorize the role of harvester in the COVID-19 period. The first of these descriptors, the RLI, is a measure of the activities in an occupation that can be performed at home, with 1 being all activities remote-friendly and 0 being no activities performable remotely (del Rio-Chanona et al., 70). The RLI for the broad agricultural industry as a whole has the