https://www.sup.org/books/anthropology/feeling-machines

Nick Seaver: Readers may have heard about care robots before, perhaps even within the context of the aging population of Japan, but your book takes a fairly unique approach to examining how care robots actually work in practice. What would you say is distinctive about the methods you use in this book, and what kinds of things did these methods orient you toward finding out?

Shawn Bender: I think you’re right to see a link between the aim of the research for the book and the methods that I used in order to carry it out. From early on, I was curious about how care robots were used in practice, particularly the care of the aged. This is not to say I had no interest in laboratory research. In fact, my first contacts in the field of care robotics were academics in Japan who were doing laboratory research on human–robot interaction. As fascinating as much of this work was, I wondered about how closely the use of robots in sites of care matched claims about the utility of these devices. What features would care workers see as vital? What would they minimize? What would leave them satisfied or dissatisfied? How would robots be treated once their novelty wore off and they came to be integrated into care routines? On the one hand, I thought that a view from the everyday might temper, or at least nuance, sometimes sensational claims I had read about the benefits of robotics for Japan’s aging society. On the other, I believed that this approach could add to work of other researchers who had introduced robots into care facilities through shorter-term field studies.

This interest in observing what happens on the ground with care robots would ultimately lead in three directions. First, it took me out of Japan once I learned that some of the earliest and most energetic adoptions of Japanese care robots were happening outside of Japan. For example, the furry seal robot called Paro, which was invented in Japan, was first integrated into dementia care in Denmark. And it was German doctors who first adopted the robotic exoskeleton HAL as a tool in the rehabilitation of walking disabilities. Following these two devices turned into fieldwork in both Denmark and Germany. The case of HAL in Germany would also lead me to expand the scope of my research to include the care of people with disabilities, in addition to older adults. Finally, and perhaps somewhat ironically given my initial pathway into the project, I realized over time that I could not so easily separate the integration of care robots in sites of care from the work of developing them technologically in the laboratory. Work with care robots in the field, I learned, was entangled with the technical refinement of robots in corporate laboratories. In sum, a project that I intended to complete in one site in one country in one year turned into a study that had me visiting multiple countries several times over a number of years.

Nick Seaver: The multisitedness and focus on actual users of these technologies certainly expands the frame and routes around some of the hype and hype-puncturing that captures much discourse on the subject. One common question that comes up in multisited projects is this: Are you studying one geographically distributed phenomenon or a collection of diverse, yet related phenomena? For instance, would you say that you encountered culturally distinct understandings of care robots in Denmark, Germany, and Japan? Or were differences more a matter of the particular institutional settings you found yourself in?

Shawn Bender: It’s certainly the case that including multiple countries in the research frame settles some questions and raises others. Commonly aired beliefs about the deep Japanese affection for robotics, for example, run into trouble when viewed against higher adoption rates elsewhere.  Similarly, the idea that Japanese prefer robotic helpers over immigrant caregivers suggests that there is some inherent incompatibility between the two. Such a claim makes little sense in Denmark, for example, where immigrant caregivers work alongside Paro and other kinds of digital technologies. Clearly, other factors are at work. Restricting one’s view to one national-cultural frame can sometimes obscure rather than enlighten.

At the same time, it really wasn’t an interest in comparison per se that drove my fieldwork in these three countries. Instead, the primary impetus was a curiosity about how people were using Japanese care robots in practice. It turned out to be that caring with Japanese robots is geographically distributed in a way that I didn’t initially anticipate. The inclusion of three countries in my study was thus incidental to the process of learning about actual care practice. But, while it was never my primary aim to get, say, a Danish view on care robotics in comparison to a Japanese one, the approaches to care taken in my field sites were certainly shaped by a range of local social, political, and cultural understandings. There were also attitudes toward caring with robots that were shared across national sites—machines do not travel without a package of ideas accompanying them. Perhaps most surprising, and somewhat deflating as an anthropologist, I found that one of the most powerful influences on whether a technology was adopted here or there had nothing to do with cultural attitudes toward technologically assisted care. Instead, it was the presence or absence of an insurance system that supported the cost of caring with new technologies, robots or otherwise.

Nick Seaver: That’s a really interesting outcome: it seems to deprioritize some of the more culturally essentialist ways of thinking about technology (especially that overloaded Japan/robots nexus) in favor of an account that’s more focused on social structures, institutions, and the distribution of resources. Do you see this as significant for the anthropology of technology more broadly? Does it tell us something about how we, as anthropologists, might want to reorient toward our objects?

Shawn Bender: You’re right that I did have to push back against assumptions about those Japanese and their robots.  Essentialism is one thing, but I would be hesitant to minimize the role of cultural understandings in shaping how technologies are designed, deployed, and given meaning in specific places. Part of this is my own bias. I’m most interested in studies of technology that strive for holism and that use fine-grained ethnographic accounts to show how material objects are embedded in social contexts. These is the kind of writing that anthropologists still do, even as we have absorbed the influence of science and technology studies and sociologists of technology—that is, fields that tend to emphasize the role of structure and institutions over norms, values, and beliefs.

I guess as someone who works primarily in Japan, my preference would be to expand what we might include as parts of the culture. Some parts of the world, like Japan, occupy a slot within anthropology in which cultural factors border on overdetermining. And it seems that there is general agreement as to what those cultural factors are (for example, emphasis on the group over the individual, or beliefs that machines have souls like humans). Other than analyses of, say, individualism or bounded personhood, there is far less of this for places in Europe or North America. Technologies from these parts of the world emerge basically culture-free. Take for example, the iPhone. There have been so many studies of smartphones and smartphone culture, but very few of them consider the iPhone itself as a product of American culture or as a distinctly American object. As a product of a unique Silicon Valley ecosystem, yes. But as a particularly American technology? I don’t think I can name one. This is not to suggest that such a study couldn’t be done. It is to say that this is not where most analysis begins or is even expected to begin. Yet, for studies of Japanese technology, culture is often the starting point (or endpoint).

At the root of this is an assumption that Japan is more isolated from global trends than other countries. In my work, however, I found that the cultural factors of most relevance were more global than national. These were connected to new norms and values regarding health and the ends of care. I’m referring to attitudes toward care that see health as more than the absence of disease and that view the goal of care as maximizing the quality of life lived, not just preserving or extending life itself. Standards of care for dementia have evolved to place greater emphasis on stress reduction, social integration, and psychological wellbeing. Interventions to manage symptoms of chronic disease and injury increasingly respond to a patient’s subjective experience of physical wellbeing. This culture shift has led to new ways of evaluating good and bad care, which in turn has led to public and private insurance schemes to support them. I think it’s hard to understand the appeal of care robots for many, but of course not all, welfare systems without understanding how this culture of health has become normalized worldwide.

Nick Seaver: For potential readers out there, I think the book does a great job of taking that holistic approach, while also following its objects transnationally. Looking forward from here: What are you working on now? Are you continuing along the lines started in this book?

Shawn Bender: Yes and no. I want to continue exploring the social effects of robotics and automation but in a different context than care. I’ve been struck by efforts over the past decade in Europe, the US, and Asia to automate more and more aspects of agricultural work. You may have heard that a number of large manufacturers have been pushing the development of fully autonomous, self-driving tractors. While the commercialization of fully autonomous tractors is unlikely to be happen, it is the case that most existing tractors already operate at least semi-autonomously. (For example, GPS guidance has been standard on most crop machines since the late 1990s.) The drive for complete autonomy in crop agriculture made me curious about existing applications of robotics, AI, data analytics, drones and crop surveillance, among other forms of automation in the agricultural sector.

I’m most interested in how increasing levels of digital automation are reshaping what it means to be a farmer. Right now, I’m staying close to home to examine how local dairy farmers who have adopted automatic milking systems think about their animals and their work, but I anticipate extending my project scope to include other forms of food production and to expand my field sites to include comparison cases abroad in Japan. Despite the drastically different sector of activity, I’ve been surprised so far at how similar concerns about an aging workforce are driving the push for automation in agriculture just as they have in the field of eldercare. This is not a project that I think I would have developed had I not done the research that went into Feeling Machines, and I look forward to applying the insights gained through that projectto this new area of study.