Dr. Karolina Zawieska

School

University College Dublin

Keywords

social robots, anthropomorphism, Human-Robot Interaction, symbolic interactionism, meaning-making

Abstract

Anthropomorphic artefacts have long been objects of scientific curiosity and artistic explorations as an illusion of life has been created in sculptures, puppets and automata. Only now, however, technology and the underlying mindset have evolved to the point that we aim to incorporate human-like machines in our everyday life, where robots would act as social actors that equal human beings. This is how the study of robotic systems becomes also the study of human beings and social interaction between the two. The main field that addresses this subject, which is also the research area this work adheres to, is Human-Robot Interaction (HRI). In line with the interdisciplinary nature of HRI research, this work addresses the HRI field from the symbolic interactionism (SI) perspective, i.e. one of the main theoretical frameworks developed in social science in general, and in sociology in particular.

The main goal for HRI research up to date is to build robots capable of communicating and interacting socially with human beings. This is why social robots resemble human beings, to a varying degree, with anthropomorphisation (attribution of human characteristics to non-human objects and entities) constituting a common frame of reference for different HRI studies. The anthropomorphic effect in robots has been generally viewed as an outcome of two elements, namely anthropomorphic robot design and anthropomorphic projections made by people, or both. This works hypotheses that HRI research tends to prioritise the role of robotic design in creating the anthropomorphic effect and reduces the role of human users to that of reacting “automatically” towards specific stimuli provided by robots. Such an approach leaves little room for the human capacity to attribute meanings to robot appearance and behaviour, which stands in a sharp contrast with symbolic interactionism. The latter emphasises the human ability to actively construct and modify meanings (use symbols) and interpret the things they deal with rather than merely react towards the external or internal stimuli that act upon them. At the same time, however, the ability for symbolic interaction seen as a uniquely human quality has been increasingly questioned within the contemporary variants of symbolic interactionism. This is part of a broader discussion on the concept of the human being developed in a particular socio-cultural context, where the difference between humans, animals and technological artefacts has been increasingly blurred. This work follows the classical version of symbolic interactionism and views anthropomorphisation of robots as a chance to better understand how we view ourselves in terms of what makes us uniquely human, as defined in both the HRI and SI research. This is also to highlight the anthropocentric perspective within and outside the HRI field, where the difference between the human being and non-humans is in essence and not only in degree.

The aim of this work is to explore how the concept of meaning-making is conceived in HRI research in the context of anthropomorphisation of robots. In order to address such a question, this study follows the qualitative theoretical and methodological perspective and employs thematic analysis to analyse the data derived from selected websites dedicated to anthropomorphic robots as well as the data collected through interviews conducted with HRI researchers. The results of the study show humans are perceived not so much as automata that respond ‘automatically’ to anthropomorphic cues as animals that react ‘unconsciously’. Also, the study leads to the conclusion the term ‘anthropomorphism’ and ‘human-likeness’ are sometimes seen as two different concepts in HRI research which has particular implications for how we conceive the role of the human actors in the process of anthropomorphisation.

In terms of practical implications for robot design, this research proposes to conceive the anthropomorphic effect as a phenomenon that relies on conscious engagement with robots rather than unconscious reactions towards anthropomorphic design. This includes exploitation of a mismatch between the anthropomorphic design and the human frame of reference rather than attempts to reduce such a mismatch. It also leaves the doors open for the development of the human-robot-human interaction paradigm that fully explores the social nature of the anthropomorphic effect.