Institut für Soziologie Using sociological role theory for investigating what we know so far about the introduction of socially intelligent robots in human work settings Martin Meister Department of Sociology Technical University Berlin Martin Meister Graz 10/05/2016 1/20 Introduction Institut für Soziologie Aim of this talk: Explore the idea to use sociological role theory to describe the sociotechnical changes that are induced by the introduction of social robots in mundane human environments Goal of conceptual proposal: To allow (other than in STS) a comparison of single case descriptions Martin Meister Graz 10/05/2016 2/20 Introduction Institut für Soziologie Outline: 1 2 3 4 5 6 7 A simple case Role theory and STS Illustration: Return to the simple case Intermezzo The case of legal ascriptions of responsibility Desiderata Conclusions Martin Meister Graz 10/05/2016 3/20 1. A simple case Institut für Soziologie Investigations of interactions with a vacuum cleaner in a household Unintended result of a study (Sung et al. 2010): The introduction of “robots changed social roles in the house. They induced collaborations among more householders to complete a manual task that used to belong to one person prior to robot use (e.g., Mom for cleaning). This finding suggests that even for a simple utilitarian robot, the ability to act term acceptance” (ibid: according to social rules can become critical for long-term 427) Introduction of the new artefact led to changes in the distribution of domestic cleaning tasks, e.g.: • Dad as technical expert (for what the robot cannot clean) becomes involved • Children have to be surveyed/convinced not to experimentally destroy the robot (throwing it from the roof/drowning it in the basin) Martin Meister Graz 10/05/2016 4/20 1. A simple case Institut für Soziologie Rearrangement of the entire domestic role-set Introduction of the robot changed • • not only some single roles but the entire system of organizational roles in the household observed. Levels of scale involved Obviously, not only the organizational level is relevant, but also • Regime-level issues, e.g. – domestic gender relations – manufacturer-consumer relations • Micro-level issues, e.g. – different (technical) cleaning capabilities of the robot initiates Dads technical interest triggering his cleaning activities – Children's diabolic experimenting (definitely human) with the robot initiates (human) educational response Martin Meister Graz 10/05/2016 5/20 2. Role theory and STS Institut für Soziologie The STS-perspective on Social Robotics Asking about the rearrangement of the entire role-set seems to be the appropriate question for studying Social Robotics • • robots built to (physically, emotionally etc.) interact with humans in their ‘natural’ surroundings not laboratories, but homes, assembly lines or nursing homes The findings of this simple case seem to be in line with STS-studies on Social Robotics • e.g. the rich observation study Sabanovic/Chang 2015 about “robot sociality as an emergent relational property of the interactions between the robot and its social context“ (ibid.: 10): These examples illustrate the ‘‘interpretive flexibility’’ of PARO’s sociality—its ability to be interpreted in different ways by different social actors and redefined based on personal needs of users and the institution it is used in” (Sabanovic/Chang 2015: 12). Martin Meister Graz 10/05/2016 6/20 2. Role theory and STS Institut für Soziologie Two shortcomings of this STS-perspective 1. Perspective is conceptually flat It remains unclear/metaphorically what terms like • • situated mutual adaption robot as social actor (Alac) etc. conceptually mean. 2. Perspective is (mostly explicitly) restricted to showing that the use of the robot shows interpretative flexibility, mutual adaption etc. for single cases. This excludes more general descriptions/conceptual efforts, which are the prerequisite for comparison and for building a state of knowledge over many case studies. This is where, in our view, sociological role theory comes into play. Martin Meister Graz 10/05/2016 7/20 2. Role theory and STS Institut für Soziologie The tradition of sociological role theory (in a nutshell) Social roles are generalized expectations about behaviors that (more or less explicitly) prescribed by position in a network of logically or historically interrelated positions. Every position can (at least in principle) be filled with any concrete person, and because of its independence from concrete incumbents positions “do not cease to exist when they become vacant” (Dahrendorf 1968 [1958]: 34). Because of the interrelatedness of these generalized expectations in a “position field” (Dahrendorf) or a “role-set” (Merton), every change of one position affects changes to all other positions/roles. Martin Meister Graz 10/05/2016 8/20 2. Role theory and STS Institut für Soziologie Application to socio-technical networks/ innovations Because positions/social roles are independent from any substantial features of concrete incumbents (e.g. being a human actor), they can also be filled out by artifacts (e.g. delegation of prescribed tasks to a machine). But because of the relational nature of positions/social roles, any change of one position constitutes the need of mutually readapting all relations of the position field/the role set. • • E.g. when delegating a role to an artifact, the entire socio-technical network has to be readapted. Sometimes, role relations are inscribed in the artifact (the user is “scripted”, Akrich), sometimes successful innovation is an emergent process of mutual re-adaption. In spite of asking about the replacement of human skill by “intelligent” machines (the old AI-Question), the focus is on how the socio-technical network changes by the introduction of new role players (robots) – and on how the old role players (humans with their division of work and their hierarchies) adapt to the new constellation. Martin Meister Graz 10/05/2016 9/20 2. Role theory and STS Institut für Soziologie Two Features of role theory beneficial for comparison • (returning to the stated two shortcomings of STS-studies on Social Robotics) 1. Role Theory can be applied on different levels of scale of sociality/ social relations: • Micro-level (roles in interaction patterns) • Meso-level (organizational roles) • Regime-level (different bodies of actors; think of Large Technical Systems) 2. The same positions/roles in single cases can be described in more concrete, case-specific terms or with more general concepts. Both features are useful for comparison of/achieving more generaled findings from the single case studies. • To illustrate this, I return to the simple case of the domestic vacuum cleaner. Martin Meister Graz 10/05/2016 10/20 3. Illustration: Return to the simple case Institut für Soziologie If we apply these two features of role theory to the case of the introduction of the robotic vacuum cleaner, on every level of scale, a more concrete (case-specific) and a more general description are apparent: Generalization High Low (casespecific) Basic patterns e.g. distribution of initiator-responderroles Domestic division of labor Division of domestic expertise Gender relations Manufacturerconsumerrelations Interaction patterns Domestic distribution of cleaning jobs XX Micro-level Meso-level Regime-level Levels of scale • • Regime-level descriptions are per se on a high level of generalization Existing Macro-level discussions are about substantial qualities of robots (RoboEthics) or humans (comparison of national robotic cultures: Western-Japan): role theory cannot be applied Martin Meister Graz 10/05/2016 11/20 4. Intermezzo Institut für Soziologie The original plan Applying this scheme on existing observations/descriptions of Social Robotics • still work in progress, but Problem encountered Existing descriptions/interpretations are 1. mostly not detailed enough for reinterpretation/ secondary analysis • e.g. for quantitative analysis: Broekens et al. (2009) concluded: “There is some qualitative evidence as well as limited quantitative evidence of the positive effects of assistive social robots with respect to the elderly. The research designs, however, are not robust enough to establish this. Confounding variables often cannot be excluded. This is partly due to the chosen research designs, but also because it is unclear what research methodology is adequate to investigate such effects” (ibid: 1) Martin Meister Graz 10/05/2016 12/20 4. Intermezzo Institut für Soziologie Existing descriptions/interpretations are 2. focusing on other research questions • e.g. the works of the group of Pfadenhauer (2014-16) on the introduction of Paro in nursing homes. Here, the focus is on changes of self-description and observables practices of nursing personal in a nursing home. They conclude that there are – up to date – no observable changes in professional selfdescriptions and underlying nursing practices (as described in other – historical – cases, e.g. Barleys seminal studies). There might be two reasons for this: – Changing professional self-descriptions and/or practices takes time (mutual adjustment of roles), so it might be simply too early to observes this. – Focus of the Sociology of Professions is only on the relation of professionals and lay people, and not on a larger socio-technical network, which may restrict the empirical viewpoint. But look at the methodological considerations Ziegler 2016 ... but of course I have a second case that fits ... Martin Meister Graz 10/05/2016 13/20 5. The case of legal ascription of responsibility Institut für Soziologie Problems for applying existing legal regulations to Social Robots Beck (2015) lists “major challenges for legislature as well as for jurisprudence 1. Wronged party (user/consumer) has to proof defect/misuse • Difficult if computer programs or interactions are complex and not reconstructable ex post 2. Producer is liable for defects of/damages caused by his products • Difficult for robots with adaptive/learning capabilities. They react in unpredictable ways – damage can result from desired behavior => They are no normal products 3. Causing party (producer or owner) is liable for damages • Difficult to identify one and only one causing party in complex constellations of development/ production => liability gap Martin Meister Graz 10/05/2016 14/20 5. The case of legal ascription of responsibility Institut für Soziologie Proposed solution: “legal person ltd.” Analogous to corporate agents: “creating a ‘tangible symbol’ for the collaboration of all the people creating and using that specific robot” (Beck 2015) and giving it by contract a legal status as the address of liability • entered as a specific robot in a public register • “bundling of all the legal responsibilities of the various parties (users, producers, sellers etc.)” • all groups involved (developers, manufacturers, users) make payments in a capital stock => Robot as “legal personhood” can be held responsible under the terms of the contract and civil suit Martin Meister Graz 10/05/2016 15/20 5. The case of legal ascription of responsibility Institut für Soziologie Such a “legal person ltd.” is an interesting new position: an artificial compound position comprising the whole bundle of legal responsibilities If we read “legal person ltd.” as a general solution for concrete problems • though many of them are anticipated only we can apply the same scheme from role theory: Generalization High Low (casespecific) Damaged party vs. damaging party (including agent that decides about these) „legal person ltd.“ as compound position Law suits Courtroom interactions Concrete problems of responsibility ascription (liability) Micro-level Meso-level „repercussions“ (Beck) of basic legal concepts and solutions XX Regime-level Levels of scale Martin Meister Graz 10/05/2016 16/20 6. Desiderata Institut für Soziologie many ... did not go into case descriptions or conceptual proposals for the micro-level • literature on interactional roles exists in Social Robotics, generalization is an issue there and for the regime level • literature on the larger socio-technical embeddedness of robots/robotics is absent from Social Robotics, a good ides seems to look at concepts of the mutual positioning of politics, economy, law, developers, user groups etc. (i.e. Large Technological Systems, Niche-Regime-Constellations/CTA) and of course more rich case studies to be filled in. Martin Meister Graz 10/05/2016 17/20 7. Conclusions Institut für Soziologie I wanted to show at least in principle • that sociological role theory can be used to describe the changes that the introduction of social robots bring about socially (changes in the position field and the related roles) • that the role-theoretical descriptions can be scaled from case-specificity to more general terms, as to allow for comparability • and that it creates sociologically richer descriptions if different levels of scale are taken into account for the same cases. Martin Meister Graz 10/05/2016 18/20 Institut für Soziologie Thank you for your attention Curious about your suggestions Martin Meister Graz 10/05/2016 19/20
© Copyright 2026 Paperzz