AI in Context: The Labor of Integrating New Technologies

Abstract

In this report, we provide new evidence and frameworks for thinking about the near-term impacts of automated and AI technologies on society.Our report demonstrates the necessity to integrate rather than deploy AI technologies and to account for how AI technologies reconfigure work practices rather than replace workers.We examine two American work environments in which AI technologies often are imagined to have great disruptive potential: agriculture and retail.Specifically, we analyze technologies being used in farm management and grocery retail, contrasting the utopian (and dystopian) visions that tend to accompany AI with the mundane realities of current automated technologies.Our report demonstrates the ways in which automated and AI technologies tend to mask the human labor that allows them to be fully integrated into a social context while profoundly changing the conditions and quality of labor that is at stake.In our discussion, we focus on the ways in which this dynamic shifts ingrained community norms, interpersonal relationships, daily routines, and skill sets.We find that although new skills are often required, they are usually unacknowledged and uncompensated.We also find that adopting new AI technologies produces economic risks that are not evenly distributed among stakeholders, wherein more vulnerable or precarious communities are exposed to greater risks and harms than those who control the design or use of AI technologies.In the agricultural context of family-owned farm management technologies, we argue for the necessity to frame the introduction of new AI technologies as being integrated rather than deployed, emphasizing that a technology must be AI in ContextData & Society it is all the more important to ensure that those who work with and alongside AI systems are adequately protected, supported, and compensated.While "transitional" or "intermediate" phases are by definition temporary, they still are actualities that affect workers in the present and need to be addressed.AI technologies hold great promise in advancing society and addressing existing problems.However, the potential benefits must not obscure the potential perils of these technologies.These perils have nothing to do with "killer robots" or the coming of "robot overlords."Instead, they will be found in the everyday structuring potentials of AI that will benefit some members of society but leave many others behind. AI in ContextData & Society ways in which these technologies mask the human labor that allows them to be fully integrated into a social context while profoundly changing the conditions and quality of labor that is at stake.In this report, we focus on two American work environments within the broad industries of agri-food: family-owned farm management and corporate grocery retail.These contexts represent different workplace dynamics and types of work, allowing us to analyze AI adoption in contrasting contexts.We develop two frames to demonstrate the necessity to integrate rather than deploy AI technologies and to account for how AI technologies reconfigure work practices rather than replace workers.These frames also serve to bring into focus the ways in which automated technologies have had significant but under-examined consequences for everyday life and formations of power.17 Remi Scmaltz, "What Is Precision Agriculture and How Is Technology Enabling It?" AgFunderNews, April 24, 2017, https://agfundernews.com/what-is-precision-agriculture.html/.18 See, for example, Lyndsey Gilpin, "How Big Data Is Going to Help Feed Nine Billion People by 2050," TechRepublic, accessed October 11, 2018, https://www.techrepublic.com/article/how-big-data-is-going-to-help-feed-9-billion-people-by-2050/.Scholar Michael Carolan has characterized discourses around big data and precision agriculture as both "anticipatory" and "moralizing," specifically in the context of climate change.