Robots and the European Public Imagination: Eurobarometer Survey Results and Methodological Issues Răzvan Rughiniș 1,a*, Cosima Rughiniș 2,b 1 University POLITEHNICA of Bucharest, Splaiul Independentei 313, Bucharest 060042, Romania 2 University of Bucharest, Schitu Măgureanu 9, Bucharest 010181, Romania a razvan.rughinis@cs.pub.ro, b cosima.rughinis@sas.unibuc.ro *Corresponding author Keywords: Robots, Survey research, Public imagination. Abstract. We analyze the Eurobarometer 77.1 / 2012 survey data on robots, and we identify three attitudinal clusters of respondents: 1) people who adopt an anthropomorphic approach, accepting robots for high- and low-empathy tasks, 2) people who have an instrumental orientation, accepting robots only for low-empathy tasks, and 3) respondents who are generally aversive towards robots. Young people, students, the higher educated, and men are more often found in the first two categories than in the third. The third attitudinal cluster is more skeptical when evaluating robots, while the first two clusters do not differ in their assessments, being equally appreciative. We evaluate critically the relevance of such survey-based classification for real-life decision situations, especially when people s answers about robots are not based on actual experiences of interaction with robots. Still, investigating the cultural construct of robot in present-day public imagination and uncovering its variability is potentially useful for technological designers, marketers, and policy-makers. 1. Introduction Robots are a vivid presence in popular culture and, increasingly, a familiar presence in workplaces and homes. Still, they remain puzzling even more so because they come in many shapes and with widely diverse functionalities, ranging from playful toys to lifesaving medical equipment and organs, or to lethal weapons [1]. Robots depicted in cinematic, literary, scientific and other reports also span a large variety of forms and roles. For most people, robots are loosely defined collections of possibilities, informed by a visual and narrative imaginary assembled from heterogeneous sources. Actual interaction with robots is a powerful experience, due to its atypical emotional and even bodily, visceral configurations [2]. People often experience robots as autonomous agents, and relate with them (quasi-)anthropomorphically. A growing field of research explores the topic of human-robot interaction (HRI), as a distinctive form of human-computer interaction. Survey research has the potential of including a relatively large number of participants and eliciting information on a variety of situations of interactions with robots, real or, more often, imagined. Given that most large-scale survey respondents are put in the situation to reflect on unfamiliar scenarios involving robots, which are presented to them in the interview situation, we can interpret survey results more as outlines of the public imagination concerning robots, rather than public opinion: the experiential basis of people s judgments of robots has remained, until now, rather thin. Previous research indicates that there are some distinctions that orient people in their survey answers. Tasks that require moral decisions, possibly with life-and-death consequences, are evaluated differently than routine activities [3] although there is no consensus on how robots should be
involved in any of them. Overall, robots are associated with cold personalities, dominated by rationality, precision and reliability, while humans are considered superior as regards intelligence, sensation, sympathy, and feelings [1]. A related distinction separates activities that require skill but no empathy from activities that involve human understanding and subjectivity. Robots are seen as particularly fit for activities involving highly specialized abilities, particularly jobs that require memorization, perceptive acuity, and service-orientation, while humans are deemed more appropriate for tasks involving artistry, evaluation, judgment and diplomacy [4]. 2. Europeans facing robots in the Eurobarometer survey We rely on the Eurobarometer 77.1 / 2012 [5] to examine robots in the public imagination of Europeans. This large scale survey includes 27 countries and a total of over 26,751 respondents. We use sampling weights to obtain results that are statistically representative for the EU27 adult population (aged 15 and more). Rather than outline dominant opinions and their covariates, we are interesting in exploring variability in how robots are imagined in the survey situation. In order to capture types of reactions, we have classified respondents in three exploratory attitudinal types, using a K-Means cluster analysis to group them according to their pragmatic orientation towards robots. The questions included in this classificatory analysis [6] refer to respondents comfort or discomfort while interacting with imagined robots in a surgery room, at work, and in the family thus including both cold, instrumental, efficiency-driven tasks and warm, empathy-rich tasks. Analysis results, presented in Table 1, indicate that we can group European respondents in three attitudinal clusters: 1) those who have an anthropomorphic representation of robots, and are largely comfortable with all types of interactions; 2) those who have an instrumental view, accepting surgery and workplace interventions but rejecting robots as caretakers for dogs or family members; 3) people who are rather aversive towards robots in all presented circumstances. Table 1. K-Means cluster analysis: Types of pragmatic imagination of robots Here is a list of things that could be done by robots. For each of them, please tell me, using a scale from 1 to 10, how you would personally feel about it. On this scale, '1' means that you would feel "totally uncomfortable" and '10' means that you would feel "totally comfortable" with this situation. Having a medical operation performed on you by a robot (average scale value) Having your dog walked by a robot (average scale value) Having a robot assist you at work (e.g.: in manufacturing) (average scale value) Having your children or elderly parents minded by a robot (average scale value) 1. Anthropomorphic Attitudinal cluster 2. Instrumental 3. Aversive 5.70 4.53 2.07 7.24 1.89 1.84 7.60 8.04 2.99 3.97 1.51 1.38 % of cases in each cluster 23.38% 40.08% 36.54% Source: Eurobarometer 77.1 / 2012, authors analysis. N (valid) = 24168 Table 2 presents socio-demographic profiles for the three respondent clusters; the most prominent differences are marked with a shaded background. As it is to be expected, respondents who declare a generalized discomfort with robots are more often older, women, and with less formal schooling. Respondents that profess generalized acceptance are, by comparison, more frequently very young,
Please tell me to what extent you agree or disagree with each of the following statements about robots. still studying or with higher education, and men. Economic status does not seem to make a difference 1. Table 2. Socio-demographic profiles of respondent types (column % per trait categories) 1. Anthropomorphic 2. Instrumental 3. Aversive Age categories 15-24 years 20.7% 12.9% 13.0% 25-39 years 28.4% 25.1% 23.2% 40-54 years 24.7% 27.6% 25.8% 55 years and older 26.1% 34.4% 38.0% Gender Masculine 57.4% 51.4% 40.1% During the last twelve months, would you say you had difficulties to pay your bills at the end of the month? Feminine 42.6% 48.6% 59.9% Most of the time 9.0% 7.9% 12.9% From time to time 28.4% 23.5% 27.6% Almost never / never 62.5% 68.6% 59.5% Age at school completion - categories 5-14 years 7.8% 8.9% 19.6% Source: Eurobarometer 77.1 / 2012, authors analysis. N (valid) = 23620 15-18 years 35.5% 40.0% 44.7% 19-23 years 28.4% 28.9% 21.4% 24 years and more 13.1% 13.4% 6.9% Still studying 15.0% 8.7% 6.8% No full-time schooling 0.2% 0.2% 0.6% Table 3. Use of robots, interest in robots and general evaluations of robots, by attitudinal cluster membership Average scores 1. Anthropomorphic 2. Instrumental 3. Aversive Has used robots at home 8% 6% 5% Has used robots at work 7% 9% 3% Please tell me whether you are very interested, moderately interested 1.83 1.85 2.14 or not at all interested in scientific discoveries and technological developments Average score (1=Very interested, 2=Moderately interested, 3=Not at all interested) Generally speaking, do you have a very positive, fairly positive, fairly negative or very negative view of robots? Average score (1=Very positive 4=Very negative) 1.90 1.97 2.54 Robots are a good thing for society, because they help 1.77 1.80 2.37 people Average score Robots steal peoples jobs Average score 2.25 2.11 1.68 Robots are necessary as they can do jobs that are too 1.53 1.46 1.94 hard or too dangerous for people Average score Robots are a form of technology that requires careful 1.62 1.45 1.61 management Average score Widespread use of robots can boost job opportunities in 2.35 2.54 2.93 the EU Average score Source: Eurobarometer 77.1 / 2012, authors analysis. N (valid) = 20359. 1 The same conclusions derive from a multinomial multivariate regression analysis including the same categorical variables.
Table 3 compares members of the three attitudinal clusters as regards actual use of robots, interest in scientific discoveries, and various evaluative opinions concerning robots. Use is uniformly low across the sample; aversive users seem to have the lowest rate of actual contact, but workplace use was expected to be less frequent since they are, on average, older and have less formal education. An interesting conclusion is that respondents with an anthropomorphic orientation do not differ from those with an instrumental orientation as regards their evaluative profile, being equally appreciative; aversive respondents are, as expected, more concerned. Countries differ considerably as regards the distribution of respondents across attitudinal clusters: for example, at the top of the hierarchy on anthropomorphic orientations we find Poland with 43.9% and Bulgaria with 41.9%, while at the lowest end of the preference spectrum Slovenia and Luxemburg include only around 12% of people in this category (see Table 4). Table 4. Country-level distribution across attitudinal profiles 1. 2. 3. Country Total Anthropomorphic Instrumental Aversive Poland 43.9% 42.1% 14.0% 100.0% Bulgaria 41.9% 35.8% 22.3% 100.0% Denmark 38.1% 45.1% 16.8% 100.0% Slovakia 31.5% 55.1% 13.5% 100.0% Czech Republic 30.3% 54.0% 15.8% 100.0% Ireland 29.5% 27.1% 43.3% 100.0% Italy 28.8% 28.7% 42.5% 100.0% Finland 27.3% 50.2% 22.5% 100.0% Romania 25.6% 14.0% 60.4% 100.0% United Kingdom 25.6% 39.5% 34.9% 100.0% Latvia 23.3% 41.9% 34.9% 100.0% Greece 22.5% 28.6% 48.9% 100.0% Austria 22.4% 32.9% 44.6% 100.0% Lithuania 21.8% 36.5% 41.8% 100.0% Estonia 21.4% 46.4% 32.1% 100.0% The Netherlands 21.2% 62.4% 16.4% 100.0% Cyprus (Republic) 21.1% 21.1% 57.9% 100.0% Malta 21.1% 26.3% 52.6% 100.0% Hungary 20.8% 40.0% 39.2% 100.0% Spain 18.9% 32.4% 48.7% 100.0% Belgium 18.3% 46.6% 35.1% 100.0% Portugal 17.3% 26.1% 56.6% 100.0% France 16.8% 44.9% 38.3% 100.0% Sweden 16.3% 69.0% 14.7% 100.0% Germany 13.7% 47.9% 38.4% 100.0% Slovenia 12.8% 57.8% 29.4% 100.0% Luxembourg 11.5% 38.5% 50.0% 100.0% Total EU27 23.4% 40.1% 36.5% 100.0% Source: Eurobarometer 77.1 / 2012, authors analysis. N (valid) = 24168.
3. Methodological issues While the large scale of the Eurobarometer sample guarantees statistical representativeness at country level and at the EU27 population level, a core methodological question concerns the ecological validity of results. More specifically, how informative are respondents answers for their beliefs, attitudes, and behaviors outside the interview situations? Are their answers grounded in stable opinions and orientations, or do they represent methodological artifacts, induced in interaction with the researcher? On the skeptical side, there are strong reasons to doubt that such survey-based typology of attitudes expressed in abstracto would remain valid if respondents were actually in the situation to decide surgery assistance or home care for specific persons, dogs, and robots. People s assessment of robots capabilities are strongly influenced by their distinctive features, including appearance and visible displays of performance [7]. Moreover, there is a wide space for situational variability in assessments, depending on the here and now of embodied, emotional interaction. From a cost-benefit perspective, decisions concerning one s health, job, or welfare for pets, children, and other significant others often bear significant consequences, including financial and time expenditures: were robots to be a cost-effective alternative, it is probable that much of the stated discomfort in survey situations would become just one consideration amongst many. This is not to say that such surveys are valueless. On the one hand, they illustrate the cultural construct of robots at a given place and time. As we have observed in Table 3, this construct draws presently much more on fictive accounts and various media reports than on actual experiences of use; still, as robots become more frequent, evaluations grounded in direct contact will grow more common. Vicarious as it may be, the current robot construct is part of the public imagination that is relevant for designers, marketers, and policy makers. Innovative technologies often challenge such received wisdom; surveys can inform the industry of the current cultural configuration of a given social environment, facilitating exploration, persuasion, and user discovery. 4. Conclusions We present a classification of European respondents according to their attitudes towards robots. We use the large-scale Eurobarometer 77.1 / 2012 survey and we identify three profiles of hypothesized use: 1) anthropomorphic approaches that accept robots in low- and high-empathy tasks, 2) instrumental approaches that only accept them in low-empathy tasks, and 3) people who are generally aversive of interacting with robots. There is some differentiation between profiles on socio-demographic indicators: young people, students, persons with higher formal schooling, and men are more often members of the first two clusters. As regards global assessments of robots, the first two types present similar average evaluations, while the third appears slightly more skeptical. We find considerable country-level heterogeneity; future research could be useful in identifying cultural events that promote robots as positive characters. While the proposed survey-based classification is probably a poor predictor of future behavior in decision-making situations, we consider it a valuable description of the cultural construct of robots in present-day public imagination, informative for technological designers, marketers, and policy-makers. 5. Acknowledgement This article has been supported by the research project Sociological imagination and disciplinary orientation in applied social research, with the financial support of ANCS / UEFISCDI with grant no. PN-II-RU-TE-2011-3-0143, contract 14/28.10.2011. Parts of this work were carried out at the EUROLAB at GESIS - Leibniz Institute for the Social Sciences, Cologne, Germany, May 2013.
References [1] K. O. Arras and D. Cerqui, Do we want to share our lives and bodies with robots? A 2000-people survey, 2005. Available: http://srl.informatik.uni-freiburg.de/publicationsdir/arrastr05.pdf [2] J. E. Young, J. Sung, A. Voida, E. Sharlin, T. Igarashi, H. I. Christensen, and R. E. Grinter, Evaluating Human-Robot Interaction. Focusing on the Holistic Interaction Experience, International Journal of Social Robotics, vol. 3, no. 1, pp. 53 67, Oct. 2010. [3] Aj. Moon, P. Danielson, and H. F. M. Loos, Survey-Based Discussions on Morally Contentious Applications of Interactive Robotics, International Journal of Social Robotics, vol. 4, no. 1, pp. 77 96, Nov. 2011. [4] L. Takayama, W. Ju, and C. Nass, Beyond dirty, dangerous and dull: What everyday people think robots should do. Proceedings of the 3rd ACM/IEEE International Conference on Human-Robot Interaction (HRI), pp. 25 32, 2008. [5] European Commision, Eurobarometer 77.1 (2012). TNS OPINION & SOCIAL (Producer), Brussels and GESIS Data Archive, Cologne, Brussels, Cologne, 2012. [6] European Commision, Eurobarometer 77.1 Basic Bilingual Questionnaire. TNS OPINION & SOCIAL (Producer), Brussels and GESIS Data Archive, Cologne, 2012. [7] M. Lohse, F. Hegel, and B. Wrede, Domestic Applications for Social Robots - an online survey on the influence of appearance and capabilities, Journal of Physical Agents, vol. 2, no. 2, pp. 21 32, 2008.