ROBOTS IN THE WILD
An Ethnography of Robot-Human Interactions outside the Laboratory
This research project looks at how robots interact with the French public in the ‘wild’; that is, outside of the controlled environment of the laboratory or the psychological experiment. The aim is to understand concrete human-robot interactions as they happen from an ethnographic perspective. Intensive participant observation at sites where robots are currently employed in Paris – at the Mairie du 15ème, for example – and interviews with those who interact with them, will be used to map the lived experience of human-robot interactions in as much detail as possible. This data will then be brought into conversation with two aspects of current discourses in and about social robotics in Japan and the EU: animism and personhood.
Engaging with the notion of robot personhood, the project looks at how the identity of the robot is understood by those who interact with it: What is its gender? Its ethnicity? What do the participants think about the robot’s abilities and limitations? How do engineers and handlers conceive of its capacities? By going beyond ideas of nationally bound robot culture, this project will contribute to an intercultural understanding of social robotics, its uses and its limits.
Contrary to the restricted setting in laboratories that many observations of Japanese robotics are based on (Robertson 2017), open situations allow for unpredictable elements. Successfully managing unpredictability is considered to be a hallmark of social interaction, an art often considered unique to humans and therefore high up on the priority list of roboticists who work on humanoid robots. The field of social robotics is interesting for anthropologists because it is laden with assumptions about the nature of human interaction (Ishiguro 2009), human specificity and the role of culture in the creation and use of technological artefacts (Kubo 2015). Assumptions concerning each one of these aspects have culturally specifically histories and were shaped by different cosmological and religious ideas in France and Japan.
The ludic automata of Jacques de Vaucanson (1709-1782) were the immediate predecessors of automation and at the same time imbued with the heretical whiff of playing God (Bertrand & Montègre 2013). Philosophical arguments about the working of the human apparatus were made manifest and rejected in the figure of Descartes’ mechanical daughter, an automaton his opponents insinuated he constructed to replaced his deceased daughter, Francine (Kang 2017). Luddite uprisings in England and later in France were based on the understanding that automation would render human labour superfluous, a fear that has resurfaced in the 21st century.
In Japan, on the other hand, dolls served as human simulacra in rituals of purification since at least the Heian period (794-1185). Human beings could be replaced by dolls and vice-versa, a logic of substitution which later found its fullest expression in the Bunraku puppet theatre. The import of Dutch clockworks after 1543 made possible the construction of moving dolls, so called Karakuri-Ningyō, that were considered luxury playthings and entertainment mechanisms (often mounted on parade floats).
While the ludic element was present in both France and Japan, the religious and cosmological underpinnings posit the automata in different registers: in the Judeo-Christian tradition, the relationship between God and his creation can be conceived of as vertical hierarchy; in the Japanese case, humans and their simulacra are said to exist on the same horizontal ontological level. While such ideas about clear-cut cultural differences inform public debates, policy discussions and projects of nation-branding (Katsuno 2015), the lived experience of human-robot interaction is rather more complicated. To begin with, Pepper, the most widely used robot in the entertainment, care and service sector, has been branded as a Japanese product, despite its French origins. Designed by roboticist Bruno Maisonnier, assembled in China and sold by the Japanese telecom giant SoftBank, Pepper has become a boundary object in its own right: a technological artefact that can adapt to the needs of local users and still maintain a recognisable identity across different national contexts.
This project examines the interactions between contemporary robots and non-specialist users. It interrogates how the legacies of previous interpretations of ludic technology influence frames of reference in the present: what do people refer to when making sense of human-robot interactions? It specifically focusses on two aspects of human-robot relations: the Japanese notion of ‘techno-animism’ (Okuno 2002) and the recent debate on robot personhood (Burdett 2020).
Techno-animism: Japanese commentators and foreign observers alike have pointed out that Japan is a culture that welcomes robots with open arms (Katsuno 2015). This openness to robots as social others is often ascribed to Shinto, the animistic indigenous religion on the Japanese archipelago (Blok & Jensen 2013). In the Shinto worldview, natural and man-made objects are inhabited by spirits; no absolute boundary is drawn between the animate and the inanimate, the human and the non-human. It follows that robots are also seen to be inhabited by spirits and thus are treated as social others rather than as alienated and alienating technology. This narrative has been used by the second Abe administration (2012-2020) to promote Japan as the coming ‘kingdom of robots’ in which the development of care robotics would ease the labour shortage in the care sector.
Robot personhood: In 2017, the EU commission on civil laws and robotics suggested in a motion that in order to clear up issues of liability, robots should be granted “electronic personalities”. While this addresses the question of legal personhood, what happens in human-robot interactions is much more complex. While American roboticists in the 90s have argued that social robots must exhibit simulated human traits to successfully interact with humans, anthropological inquiries into anthropomorphism have shown that the attribution of personhood has a much lower threshold: Alfred Gell has shown how icons and aniconic representations in religious contexts were given personhood by worshippers despite not exhibiting human-like behaviour (1998). In a seminal paper on the possibilities of a comparison between robots and deities in a Himalayan context, Denis Vidal has pointed out that the identification of the social other as possessing human capacities was not necessary for meaningful exchanges to take place, nor did the social other need human-like traits, a phenomenon he calls sub-anthropomorphism (2007). Furthering this line of inquiry, the French anthropologist Emmanuel Grimaud (2012) described attribution of life as first and foremost the concern of the creators of robots, who try to create a moment of “ontological confusion” between people and robots. He argues that on the side of the users, this confusion is resolved very quickly, but this does not stop the interaction: the initial “ontological confusion” is replaced by an “ontological gradation”, in which “para-humans” – everything that fosters attachment and connectivity, in Sherry Turkle’s words “relational artifacts” (2006) – become partially animated in “a series of larger or smaller shifts, analogies, confusions and alterations” (“toute une série de glissements, d’analogies, de confusions et d’altérations plus ou moins grandes”) (Grimaud 2012: 91).
Research Scope and Methodology
The proposed method is ethnographic fieldwork in several sites in which robots are currently used in Paris and its environs. The companion android Pepper is currently in use in different locations in Paris: The Mairie du 15ème has both Pepper and Nao working as receptionists and as a first point of contact between citizens and local government. The company Hoomano who coordinates the software development for the interactions in based in Tokyo, but has the only other regional office in Paris. Génération Robots, the company that is tasked with the maintenance of Pepper and Nao is based on the outskirts of Paris.
The research protocol includes participant observation of human-robot interactions, interviews with those who interact with the robot, but also with other observers, those who maintain and repair it, and those decision-makers responsible for the introduction of robots into concrete environments. If possible, a set of ethnomethodological experiments will be conducted: their aim is to deliberately disturb the smooth flow of social interaction and to observe how the different actors react to the disturbance and how they try to ‘mend’ the interaction. This could take a number of forms: what happens when someone starts to shout at a robot, for example? How do others in the vicinity react? Three months of intensive daily fieldwork should suffice to produce a suitable amount of data. Video recordings (with permission) of human robot interactions will be taken for later analysis and illustration of the fieldwork material.
By documenting and analysing how robots interact with the French public in different contexts, the study aims to critically interrogate bodies of knowledge concerning the nature of the robot and to provide evidence for a diversity of possible relationships. How do people make sense of encountering a new being? From this core question, three inquiries are pursued:
- Is the robot perceived to be ‘alive’? Under what conditions do users experience it as an animated entity? What capacities are associated with animation? How is this perception fostered by handlers who demonstrate how to interact with the robot ‘correctly’? What are the normative dimensions of these interactions?
- Do users, visitors, observers and maintenance workers attribute gender to the robot? What interactions facilitate such an attribution? What are the limits of anthropomorphism at work in these interactions?
- As a new being in their environment, how do users conceive of the origins of the robot? Is the attribution of a place of origin or a “made in” label conducive to the interactions that follow? Does the robot have an ‘ethnicity’ in this popular perception?
Rather than to think of techno-animism as a pre-formulated belief system, the hypothesis guiding this inquiry is that the life, gender and ethnicity attributed to the robot must be understood as a result of the concrete relationships that it enters into. In other words, techno-animism and ideas of robot personhood are not necessarily present at the beginning of the interactions between human and robots, but result from them. This would lead away from a culturally-bound paradigm of Japanese social robotics, towards a more open-ended, intercultural understanding of human-machine interactions. In sum, the aim of this study is to contribute to the project of a comparative anthropology of artificial life forms (Grimaud & Vidal 2012).
Bertrand, Gilles and Gilles Montègre. ‘De la curiosité pour les automates dans l’Europe des Lumières. Machines et mécanismes en Italie au prisme du regard des étrangers (1680-1820)’. In L’Automate, modèle, métaphore, machine, merveille, edited by Aurelia Gaillard, Jean-Yves Goffi, Bernard Roukhomovsky, and Sophie Roux, 429-457. Pessac : Presses Universitaires de Bordeaux, 2013.
Burdett, Michael S. ‘Personhood and Creation in and Age of Robots and AI: Can We Say “You” to Artifacts?’ Zygon: Journal of Religion and Science 55, no. 2 (2020): 347–60.
Gell, Alfred. Art and Agency: An Anthropological Theory. Oxford; New York: Clarendon Press, 1998.
Grimaud, Emmanuel. ‘Androïde cherche humain pour contact électrique: Les cinétiques de l’attachement en robotique. Gradhiva: Revue d'anthropologie et d'histoire des arts 15 (2012) : 77-101.
Grimaud, Emmanuel and Denis Vidal. ‘Aux Frontières de l’humain : Pour Une Anthropologie Comparée Des Créatures Artificielles’. Gradhiva: Revue d'anthropologie et d'histoire des arts 15 (2012): 5–25.
Ishiguro, Hiroshi. Robotto to Wa Nanika? Hito No Kokoro Wo Utsusu Kagami [ロボットとは何かー人の心を映す鏡] (What Is a Robot? A Mirror to Reflect the Human Mind/Heart) . Tokyo: Kōdansha, 2009.
Jensen, Casper Bruun, and Anders Blok. ‘Techno-Animism in Japan: Shinto Cosmograms, Actor-Network Theory, and the Enabling Power of Non-Human Agencies’. Theory, Culture & Society 30, no. 2 (2013): 84–115.
Kang, Minsoo. The mechanical daughter of René Descartes: The Origin and History of an Intellectual fable. Modern Intellectual History 14(3) (2017), 633-660.
Katsuno, Hirofumi. ‘Branding Humanoid Japan’. In Assembling Japan: Modernity, Technology and Global Culture, edited by Dolores P. Martinez, Merry White, and Griseldis Kirsch, 205–30. Berne: Peter Lang, 2015.
Kubo, Akinori. Robotto No Jinruigaku: Nijūseki Nihon No Kikai to Ningen [ロボットの人類学ー二十世紀日本の機械と人間] (Anthropology of Robots: Machines and Humans in 20th Century Japan) . Kyoto: Sekai-shisō-sha, 2015.
Okuno, Takuji. Ningen, Dōbutsu, Kikai – Tekuno-animizumu [人間・動物・機械ーテクノ・アニミズム] (Human, Animal, Machine: Techno-Animism) . Tokyo: Kadokawa Shinsho, 2002.
Robertson, Jennifer. Robo Sapiens Japanicus: Robots, Gender, Family and the Japanese Nation. Berkeley, CA: University of California Press, 2017.
Turkle, Sherry. ‘Relational Artifacts with Children and Elders: The Complexities of Cybercompanionship’. Connection Science 18(4) (2006): 347-361.
Vidal, Denis. ‘Anthropomorphism or Sub-Anthropomorphism? An Anthropological Approach to Gods and Robots’. JRAI 13 (2007): 917–33.