Emerging Technologies at Work: Policy Ideas to Address Negative Consequences for Work,Workers,and Society

Many and Varied Data-Driven or Intelligent Technologies

Although attention to emerging technologies has focused on AI and algorithms that process digital data, both of which are incorporeal in that they have no physical embodiment, some emerging technologies do take physical form and interact with humans or physical objects. In addition, some blend together incorporeal and physical components. This difference in form is one factor that may play into workers’ sense-making abilities with respect to a new technology to the extent that the form makes transparent (or not) the activities and logic by which the technology operates. Table 1 provides descriptions and examples of each type of technology based on its form (incorporeal, physical, or blended), with details in the sections below.

OPEN IN VIEWER

Table 1.

Descriptions and Examples of Emerging Incorporeal, Physical, and Blended Technologies

	Type of technology
	Incorporeal	Physical	Blended
Description	Having no physical embodiment; workers engage with them through a computer or mobile device interface as a substitute for face-to-face human interaction or physical machine interaction	Embodied; often engaged in physical tasks with humans or physical objects	Union of incorporeal and physical technical components
Examples	•  Digital platforms •  Facial recognition software •  Judicial sentencing and parole risk assessment software •  Hiring software •  Credit risk software	•  Robots •  3D printing •  Drones •  Exoskeletons	•  Augmented reality •  Virtual reality

Technological Suites with a Multiplicity of Functions

Because so many emerging technologies are digital technologies, interlinking them as components in large technological suites is increasingly possible, with consequences for work, workers, and society. Consider remote work. Prompted in the 1970s by the oil crisis, “telework,” as it was first called, relied on workers substituting their daily gas-consuming commute to the office with work from home or nearby centers, communicating with the office by telephone. In the 1980s, new technologies such as desktop computers, in combination with file transfer protocols that enabled file sharing, supported an expansion of the kinds of work that could be done remotely. By the mid-1990s, the emergence of the internet and the growing digitization of work artifacts made working remotely even easier, but still managers were not allowing nor were workers embracing remote work at predicted levels. ³ By 2005, estimates of remote workers were at 1.8 million people out of a working population of 149.3 million, or roughly 1% of workers (Lee and Mather 2008; Allen, Golden, and Shockley 2015).

By about that same time, however, managers had come to realize that remote work need not be performed by former commuters now working in their homes; rather, the growing suite of interlinked technologies used to support office, service, and professional work meant that managers could just as easily hire workers in a distant city or country. The outsourcing and offshoring of work thus emerged as an effective cost-cutting labor strategy, typically resulting in lower wages, benefits, and job security for workers in both outsourced domestic locations as well as offshore locations. The call-center industry (Batt, Holman, and Holtgrewe 2009) and engineering design and analysis in the automotive industry (Bailey, Leonardi, and Barley 2012) are prime examples of how outsourcing and offshoring grew with these technologies.

The total numbers of outsourced and offshored jobs across all industries can be as difficult to calculate as the number of remote workers. Kimball and Scott (2014) calculated that 3.2 million US jobs (75% of which were in manufacturing) had been lost just to China since 2001. Meanwhile, individualized remote work among domestic office workers in the United States began to see a resurgence in the early 2000s and soared with the COVID-19 pandemic when, at its height, over a third of the workforce was working from home. ⁴ Technologies such as Zoom (for video meetings) and Slack (an instant-messaging system with dedicated topic, project, and group chat rooms) facilitate the coordination of these workers’ efforts in their teams, departments, and firms.

With so much work accomplished across computer networks, opportunities for leveraging new worker surveillance digital technologies (colloquially termed “bossware” or “tattleware”) have increased. For example, ActivTrak, a major vendor, had five clients in early 2020, a number that grew swiftly to 800 in March 2020, and that reached 9,000 by September 2021. ⁵ These technologies variously allow firms to count employees’ keystrokes, observe workers’ web browsing, monitor workers’ email conversations, calculate metrics for things such as workers’ email response time and network connectivity, tap into workers’ mobile calls, and monitor workers’ attention during meetings.

That work is more visible when technologies are interlinked is not new: Beginning in the 1990s, the linking of technologies within business management software systems enabled firms to look deeper into the performance of their departments (using enterprise resource planning systems) and their suppliers (using supply chain management systems). What is novel about today’s emerging technologies is that this level of panoptic oversight of how work is carried out is being implemented at the level of individual workers at nearly all rungs of the corporate ladder and in a variety of workplaces, including offices, warehouses, distribution centers, factories, and small businesses (Zuboff 2019). For example, hospital managers take advantage of the panoptic affordances of medical decision support systems to monitor and measure doctor’s efficiency in matters such as closing medical charts after patient visits and treatment (Sholler 2020). This trend provides evidence that electronic surveillance now extends up the occupational hierarchy to high-skilled knowledge workers who have never in the past been so highly monitored or seen their professional autonomy and decision-making independence so thoroughly invaded.

In short, we see that interlinked technologies can carry out multiple functions. Examples include medical decision support systems intended for diagnosis and treatment that also support the monitoring of doctors and technologies in remote work that permit both collaboration among (through sharing files, holding virtual meetings, and sending written messages) and monitoring of (through opaque tracking and observation techniques) far-flung workers. That data collected for one purpose can be deployed for other, possibly unimagined purposes is an example of “function creep” (Fox 2001: 261) and hints at the kinds of problems that data-driven systems in the workplace increasingly pose for workers across skill hierarchies, occupations, and industries.

Mass Unemployment

Although the specter of mass unemployment at the hands of AI has eased a bit since the swell of books proclaiming as much appeared in the mid-2010s, ⁶ the fear has not disappeared. Andrew Yang’s 2020 presidential candidacy, for which the ideas of tremendous technological unemployment served as bedrock, is testament to the persistence of this potential implication (see Yang 2018), but the failure of AI to achieve its developers’ aggrandized claims has directly contributed to its calming. IBM’s question-answering system Watson provides a good illustration of the dashed expectations around AI: When Watson beat human champions on Jeopardy! in 2011, anticipation of its applications in industry was high, but Watson struggled to make the switch from game show to the workplace (Lohr 2021). For example, although Watson’s developers thought that it would make considerable strides in medicine, including in cancer diagnosis (a field with an excess of data that Watson’s AI engine could mine for insights), the data proved too complex, messy, incomplete, and unreadable for Watson. Undaunted by this failure, development in AI continues, with private worldwide funding of AI firms up from $282 million in the year that Watson won at Jeopardy! to $16.5 billion in 2021. ⁷ If funding is predictive of technological advance, then mass unemployment may simply come later than originally projected.

Precarity and Changing Social Contracts

When working through digital home platforms such as ride-hailing and service-hiring platforms, workers perform as independent contractors, not employees, who interact with clients through the platform and operate out of their own vehicle, office, or shop (Ravenelle 2019). Whereas AI ushered in concerns about mass unemployment, digital platforms have added to the erosion of the social contract between employers and workers that outsourcing prompted 20 years prior (Barley and Kunda 2004). In so doing, they extend the casualization of labor and the precarious employment conditions that began in the 1970s with the disappearance of good-paying industrial jobs and that have continued steadily since (Standing 2011; Kalleberg 2013; Pugh 2015). While the concerns of mass unemployment may rest primarily in the future, those of precarity and changing social contracts reside firmly in the present.

Algorithmic Management

Algorithms are the engines of digital platforms that take on matching service providers to customers without the involvement of human managers who might negotiate, assign, and oversee the job, and hence, render workers as independent contractors. Platform workers experience a highly onerous form of management because they find themselves with few avenues for interpreting customer feedback; controlling their work experience; seeking advice or counsel; and/or negotiating for opportunities to gain new skills and knowledge, vary their experience, or shift and expand their role and responsibilities (Rosenblat and Stark 2016; Maffie 2020a; Cameron 2022; Cameron and Rahman 2022). Put simply, they work at the mercy of an algorithm whose inner workings are typically opaque to them and whose determinations could lead to suspension from the platform (e.g., following negative customer evaluations). Workers may have little recourse beyond shifting to platforms with more favorable working conditions, for example, platforms that provide greater protection against worker complaints (Maffie 2020b). Algorithms are employed in many work contexts beyond digital platforms, including as the basis for software and decision support systems in medicine, policing, judicial sentencing, hiring, and credit risk assessment. The opacity of the algorithms in these systems has sparked considerable concern as workers and others who are negatively impacted by system outcomes cannot detail explicitly how discrimination or bias occurred (Ananny and Crawford 2018).

Worker Surveillance

In the past, when people worked in factories or offices doing often physical tasks (e.g., attaching parts on an assembly line or filing papers in a cabinet), managers and supervisors kept track of performance primarily through personal observation of workers and their output. As work became increasingly digital across occupations and workplaces, so did the measures of managerial control (Zuboff 1988). Today, managers’ technological arsenal for surveilling workers in pursuit of control over their performance has grown significantly and has expanded to include virtually all occupations, from the lowest skilled to the highest. For example, swipe cards record workers’ comings and goings in buildings while “bossware” software programs similarly track their online movements and activities. With constant surveillance comes a real and continual threat of sanction and dismissal, raising anxiety and stress among workers (Carayon 1993; Batt 2015).

Surveillance takes on a personal and even more intrusive tone in the case of wearable technologies that monitor workers’ health and safety through sensors that collect physiological and psychological data. Companies may employ digital health technologies to promote healthy behaviors via interventions or gamification (e.g., motivating employees to walk 10,000 steps daily; see Singh et al. 2015) or to assess workers’ mood (e.g., happiness, boredom, sadness, or anger; see Zenonos et al. 2016). Detractors view these technologies as emblematic of “corporeal capitalism” (Moore 2015: 8) and contend they raise significant privacy and social justice issues, including hiring and firing discrimination based on employers’ access to medical information (Nikayin, Heikkilä, De Reuver, and Solaimani 2014).

A further concern with heightened surveillance capabilities is the extra-organizational control of workers, as occurs on digital work platforms (Rahman and Valentine 2021). In a study of how “the crowd” holds service organizations accountable through social media posts, Karunakaran, Orlikowski, and Scott (2022) documented that when people berate 911 call dispatchers on Twitter or call out hotel staff on TripAdvisor, these back-stage workers who typically toil in relative obscurity are now named and shamed in public forums, providing a form of customer-achieved control that is both constant and unpredictable. Workers never know which customer interaction throughout their shift will prompt a tweet or review. The result for workers is that they may be held to standards they did not consent to and cannot negotiate. They may receive feedback or punishment that did not originate through the standard channels and that they therefore cannot mitigate, protest, appeal, or otherwise act upon through established procedures. And workers may be ostracized and judged with limited opportunity to clear their name or protect their privacy. Finally, and perhaps most grimly, Anteby and Chan (2018) argued that workers’ efforts to maintain invisibility in the face of surveillance technology provides management with justification to increase surveillance measures, ensuring a vicious cycle of rising surveillance.

Metrics and the Quantification of Performance

Quantification of outcomes has long been the mainstay of performance measurement, evaluation, and reward or punishment. The performance of salespeople is measured in the number of sales by week, month, quarter, or year; manufacturing workers by the number of parts assembled; elementary school teachers by the number of students taught and the percentage of students that pass standardized tests or progress to the next grade; and so on. The data that feature prominently and abundantly in emerging technologies, however, lend themselves to computation and calculation of metrics for work performance measurement at a much finer-grained scale than those of the past, resulting in a plethora of new metrics. For example, onboard tracking systems monitor truck drivers’ speed, routes, breaks, idling time, and other driving behaviors and practices (Levy 2015); software counts keystrokes to measure programmers’ productivity (Batt 2015); clicks to online news stories are counted to evaluate the performance of journalists (Christin 2018); and radio frequency identification (RFID) tags on garments prompt sewing factory operators, who can see their performance score in real time, to gamify their own work (Ranganathan and Benson 2020). Observers caution that the outcome of quantification of performance at work may include increased precarity, work intensification, taxing work scheduling and routing, anxiety, exhaustion, reduced job satisfaction and autonomy, stress from competition among workers, high turnover, and even efforts to align workers’ energy to achieve social harmonization and higher productivity (Moore and Robinson 2016; O’Neill 2016, 2017).

Work Group Dynamics

Robots, once something we expected to see working without human intervention on the manufacturing floor or traveling a solitary path in a distribution center, are increasingly working in consort with humans, as in the case of robotic surgery (Beane 2019; Sergeeva et al. 2020). Their co-working presence with humans causes us to reconfigure ideas of trust, work practices, and roles in work groups (Barrett, Oborn, Orlikowski, and Yates 2012; De Visser et al. 2020). These concerns are also widespread in work situations where AI is employed, with workers having to come to terms with algorithmic opacity and the autonomy of non-human co-workers, learning to trust decisions whose making they cannot trace or understand, and having to build shared mental models of how and why technologies act as they do (Stubbs, Hinds, and Wettergreen 2007; Glikson and Woolley 2020). In this sense, physical and blended technologies may fare a bit better than purely incorporeal ones to the extent that their actions are visible and transparent to their human co-workers (Wortham, Theodorou, and Bryson 2017; Ananny and Crawford 2018).

Worker Safety and Health

Popular media and academic literature are replete with accounts of issues of worker safety and health in the face of emerging technologies. For example, because digital platforms reclassify service providers as independent contractors, these workers lack the kinds of legal and other protections that organizations typically provide. Ride-hailing drivers may find themselves inadvertently chauffeuring drug dealers making their rounds; personal cooks may be witness to domestic violence; errand runners may become drug mules or money launderers; and home renters may end up hosting illicit parties and activities (Ravenelle 2019). Workers in call-center jobs face emotional exhaustion not only from customer interactions but also from constant e-monitoring of their performance (O’Brady and Doellgast 2021). Military personnel carrying out remote drone operations in foreign lands from military bases in the United States may suffer emotional and mental anguish as they consider the morality of their work, switching within the course of each working day from the hard realities of warfare to the everyday domesticity of family home life (Rauch and Ansari 2022). Workers in industries such as hospitality, retail, and food services may find they experience more stress and family conflict when they work according to hours determined by automated scheduling software; moreover, the schedules they receive make it difficult for them to seek medical treatment because they cannot make and keep appointments (Golden 2015; Alvarez, Loustaunau, Petrucci, and Scott 2020). Although the idea that the introduction of new workplace technologies may spawn new concerns for workers’ health and safety is not novel, the variety, pervasiveness, and distinctiveness of these new concerns warrant their inclusion on this list.

Broader Societal Concerns

Several potential consequences of the use of emerging technologies extend well past the workplace, affecting workers and others as members of society. Among these, the one receiving the most attention is that emerging workplace technologies may have ramifications for privacy, surveillance, and fairness in the general population. For example, when the LAPD employed Palantir Technologies algorithmic software to aid them in fighting crime, the software was linked to the databases of Papa John’s and Pizza Hut to gain names, addresses, and phone numbers (Brayne 2017).

Discrimination is a second potential consequence in home-renting digital platforms when hosts use information about prospective guests, such as their names, to screen along racial and other lines (Ravenelle 2019). Numerous studies point to how short-term rental digital platforms have transformed urban housing, with large cities such as New York, Los Angeles, San Francisco, Vancouver, Sydney, and Barcelona seeing a large percentage of long-term rental vacancies shifting to short-term ones, with consequences related to gentrification, housing shortages, homelessness, and encroachment by tourists on residential neighborhoods and their businesses (Lee 2016; Gurran and Phibbs 2017; Wachsmuth and Weisler 2018). Moreover, because wealthy people have more capital to invest in rental property, these platforms reinforce structures of income inequality by creating entry barriers for low-income people (Mermet 2021). As a final irony, among the homeless may be digital platform workers themselves (Green and Levin 2017).

A third potential consequence arises in conjunction with the possibility of mass AI-induced unemployment: To the extent that AI takes over more and more tasks and jobs in coming decades, more people will find themselves at home and out of work. Putting aside issues of how they will survive financially, there is the practical problem of how they will live their transformed lives. Here, the pandemic provides early instruction: Although divorce rates were low during the pandemic, experts think the cause was economic (divorce is expensive) and that when the economy rebounds, the divorce rate, which was at a 50-year low prior to the pandemic, will spike. ⁸ All that time spent together in the confines of the home put stress on couples and families, a situation that may arise if people are again together in the home for long periods of time due to unemployment. Additionally, as people find themselves with more time for leisure, they may put a strain on recreation centers, sports fields, gyms, neighborhood parks, and more. During the pandemic, which again is instructive, national parks found themselves overcrowded and overwhelmed with litter as people turned to outdoor recreation. ⁹

While some types of crime, such as home burglaries, may decrease if more people are at home instead of the workplace, other crimes may increase, including domestic violence, drug deaths, street crime, and cybercrime (e.g., Boserup, McKenney, and Elkbuli 2020; Miller and Blumstein 2020). Even prior to the pandemic, Case and Deaton (2015) documented “deaths of despair” from alcohol, drugs, and suicide rising among working-class Americans whose social connections were eroding, including those to employers, co-workers, and labor unions that slipped away when the social contract of employment withered under new work and organizational structures. If new ties formed in neighborhoods and communities do not arise in the wake of increased unemployment caused by the implementation and use of emerging technologies, then such health complications may continue and potentially increase, especially if combined with a rise in divorce, with ramifications for family members who may need to shoulder additional responsibilities and possibly health care costs in addition to coping with their grief. This outcome seems likely given that alcoholism and drug overdose deaths increased during the pandemic ¹⁰ just as they do after retirement (Bacharach, Bamberger, Sonnenstuhl, and Vashdi 2008; Zins et al. 2011; Grossman, Benjamin-Neelon, and Sonnenschein 2020).

In sum, a world without work, although often lauded in theory, will likely bring many changes, some of which may resonate far beyond the workplace where emerging technologies are implemented. To address these and other negative consequences, we need new policy ideas.

Promoting Worker and Society Involvement in Technology Selection, Development, Design, Implementation, and Use

With so much at stake, the process of how technology arrives in the workplace can no longer be left solely to a powerful few in technology, management, and government positions. Rather, society as a whole and workers in particular need people from many disciplines, occupations, and roles to shape technology decisions as they participate in the process. To date, such involvement has primarily occurred in the technology design phase under the banner of participatory, democratic, or values-based design (Greenbaum 1993; Friedman, Kahn, and Borning 2006).

Participation may also arise through unions that intercede to change how technology is implemented and used. For example, O’Brady and Doellgast’s (2021) work on call centers provides strong evidence of the effect of union interventions through collective bargaining and activism to reduce the emotional exhaustion associated with employer use of e-monitoring; rather than calling for a ban of the technology, unions promoted fair and developmental monitoring practices. Such union involvement has historically been most prevalent in Europe, where unions have long participated in public and firm-level policy debates to influence technology’s role in the workplace (Fröhlich and Krieger 1990).

Wajcman (2015) underscored how important policies related to technology implementation and use are by reminding us that workplace technologies do not determine social outcomes surrounding them; rather, our practices of technology use do. We find a good example of the importance of workers’ choice and agency in establishing these practices in Litwin’s (2011) study of the implementation of patient scheduling software at a large HMO: Use of the new technology was associated with increased performance and, of note, the effects were greater at clinics in which employee involvement in how to implement the software was higher. Nonetheless, in the case of emerging technologies, waiting until a technology is introduced in the workplace may not always be the wisest route. A safer approach is to ensure that broad participation occurs even before there is a technology to tailor.

Mandating Experimentation

Before motor vehicles can be put on the road, they must be tested. At the producer level, the National Highway Traffic Safety Administration regulates the safety of motor vehicles through crash testing; at the owner level, states operate inspection programs to ensure long-term safety. Many other consumer products fall under the jurisdiction of the US Consumer Product Safety Commission, which oversees testing, certification, and compliance of products in the public’s interest under federal law. Such protections for vehicles and consumer products unfortunately do not extend neatly to workplace technologies.

Currently, worker health and safety are regulated through federal and state laws aimed at ensuring that private employers protect the health and safety of workers and their families. The Occupational and Safety Health Act, like many such regulations, is primarily reactive in that it allows the government to respond to complaints (for example, by inspecting workplaces and issuing fines). A more proactive approach occurs through the National Institute for Occupational Safety and Health (NIOSH), whose many research programs investigate such matters as worker fatigue, respiratory health, hearing loss, and traumatic injury. ¹¹ When NIOSH investigations center on technology, they often do so in the context of employing it to address these other concerns. For example, research programs on wearable sensors, engineering controls, and personal protective equipment all have the aim of detecting and preventing workers’ exposure to hazardous materials. A program on occupational robotics (which includes mobile robots, exoskeletons, drones, and autonomous vehicles) more directly assesses the potential risks of workers’ use of emerging technologies as opposed to employing technologies to address other health concerns. Notably, however, the NIOSH research program on surveillance focuses not on the potential negative worker outcomes of being surveilled, but instead on using surveillance technologies to improve worker safety and health (for example, through tracking worker absenteeism).

A more concerted effort involving mandated experimentation to test workplace technologies in a manner like that conducted for motor vehicles and consumer products would signal to equipment vendors and organizations that they must consider—and are responsible for—the direct and indirect outcomes of workplace technology use. It would also provide some measure of assurance that otherwise unpredicted, unintended, and undesired outcomes of such use might be identified as early as possible and mitigated. Calls for experimentation of social technologies such as social media and search engines (e.g., Ko, Mou, and Matias 2016) point out that online platforms already carry out many large-scale experiments to guide product development (estimated, for example, as 300 experiments per day just at Bing, a Microsoft search engine; see Kohavi et al. 2013). “Responsible innovation,” characterized by continuous testing of social technologies for safety and efficacy (Meyer 2015: 100), seems equally applicable to all emerging technologies in the workplace whether incorporeal, physical, or blended, with transparency of results informing discussion of appropriate implementation and use among employers, workers, unions, and society at large.

Demanding Workforce Impact Reports

In the 1970s, the oil-spill deaths of mammals and birds along the California coastline spawned an environmental movement that forced corporations to submit environmental impact reports for newly planned development. In a similar manner, we might demand that corporations submit workforce impact studies before emerging technologies can enter the workplace. Some models already exist that could serve as a guide for such reports. In one such example, economist Alan Blinder (2009) analyzed Department of Labor data to predict which occupations may be prone to offshoring. Similar analyses might predict changes in the wake of emerging technologies’ use. By looking closely at the demands and skill requirements of potentially affected occupations, such reports could predict, for example, how many commercial drivers might lose their jobs to driverless vehicles (see Viscelli 2018). With the predictions of these reports in hand, we might make choices with how, when, and where to use emerging technologies to preserve workers’ autonomy, self-esteem, and jobs or, alternatively, we might set into early motion plans for how to create new jobs to allow for shifts in skill needs as work continues to evolve. The main point is that workforce impact reports would steer us toward more proactive rather than reactive stances toward emerging technologies in the workplace.

Creating a Governmental, Institutional Infrastructure for Technology Planning

Charting our technological future requires leadership and coordination at every level of government. At the national level, we should support and expand the efforts of the US Government Accountability Office’s Science, Technology Assessment, and Analytics (STAA) team. ¹² This team, which has taken over the duties once performed by the now-defunct Office of Technology Assessment (1974–1995), is forming external advisory boards to complement the expertise of its scientific and engineering staff. Expanding the membership of these boards to include people who through their roles are knowledgeable about our social fabric, built environment, governing institutions, and the like is essential for seriously considering the broader societal implications of workplace use of emerging technologies.

Others have called for even broader participation in technology assessment through the inclusion of everyday citizens, science museums, policy organizations, and universities (Sclove 2010). Another avenue for public involvement in government infrastructures for technology planning is the White House Office of Science and Technology Policy’s new initiative, called the “Bill of Rights for an Automated Society,” which seeks to involve “a wide array of stakeholders—in industry, academia, government, civil society, and the general public—in a national endeavor to make sure new and emerging data-driven technologies abide by the enduring values of American democracy.” ¹³ In addition, we must respond to the call to nominate members to entities such as the National Artificial Intelligence Advisory Committee, announced during fall 2021 as a mechanism for advising the US president on AI topics. At the state level, we need agencies that can translate national planning efforts into state policies and actions tailored to state industry and workforce profiles. At the local level, city and regional committees could home in on the details of urban planning, community resources, and related infrastructural work.

Spreading the Technology Base Geographically

In the wake of the pandemic, when many large technology companies allowed their employees in technology hubs such as the Silicon Valley, New York, and Boston to work remotely from anywhere, engineers, software programmers, and other technology professionals made an exodus to other areas of the United States. Data from LinkedIn suggest that their movement may have been more dispersed than the popular narrative portrayed, with cities such as Madison, Wisconsin, Sacramento, California, and Richmond, Virginia, experiencing greater percentage gains than the much-heralded gains in Austin, Texas, and Miami, Florida (Kantrowitz 2020). At least three reasons support why we might wish to spread the base of technology talent, resources, and infrastructure beyond existing technology hubs as we consider the potential negative consequences of emerging technologies.

First, a presumably sure way to reduce the “design-use” gap (Suchman 2002) that occurs when designed products prove difficult to use or have unanticipated negative consequences is for designers to live and work in the context of use. Second, as emerging technologies appear more frequently in remote and rural areas, as is the case with much of digital agriculture, it makes sense to have technical personnel nearby to decrease the time and cost of maintenance and repair of the technologies. Third, the existence of these jobs would provide high-wage skilled employment for the rural population, which in turn would help reduce the design-use gap. In sum, spreading the base geographically offers opportunities for involving the technical population in issues of implementation and use that before they may not have witnessed. The rural population will have similar opportunities for more involvement in issues of development, selection, and design—areas in which they were previously underrepresented. The result should be technology decisions that reflect a richer conceptualization of consequences and a better set of outcomes than what is achieved with the concentration of technical talent in a few large cities.