An investigation of the effects of anthropomorphism in collective human-machine decision-making

Titelangaben

André, Elisabeth ; Gimpel, Henner ; Olenberger, Christian:
An investigation of the effects of anthropomorphism in collective human-machine decision-making.
In: Proceedings of the 26th European Conference on Information Systems (ECIS). - Portsmouth, UK , 2018 . - .

Volltext

Link zum Volltext (externe URL):

Abstract

Anthropomorphism describes the attribution of human-like physical or non-physical features, behavior, emotions, characteristics and attributes to a non-human (Epley et al. 2007). The hu-man tendency to humanize (socio-)technical systems can be used in the development of anthro-pomorphic information systems (IS) to reduce emotional distance to the IS and to create a natu-ral connection between human beings and (socio-)technical systems or its components (Epley et al. 2007; Pfeuffer et al. 2018). In particular, new technologies make it possible to implement increasingly human-like features that further increase familiarization with IS. Increasing cogni-tive and emotional intelligence, contemporary and avant-garde interface design contribute to perceived human-likeness and anthropomorphism.
By enhancing IS with such complex anthropomorphic cues, it is also possible to develop in-creasingly advanced user assistance systems that adapt to the current context and the needs of their users. Advanced User Assistance Systems (AUAS; the acronym is used for both the singu-lar and plural) are IS that support users in fulfilling a task by not only offering advice on a topic, but also referring to the user's current activities and environmental conditions in order to provide context-related recommendations and advance interaction between users and with the IS (Mäd-che et al. 2016). Based on this technological progress, we increasingly see groups of both hu-mans and AUAS interacting in a collectively intelligent way (Gimpel 2015). In such collectively intelligent group decision-making settings, information and communication technologies increas-ingly do not only take the role of merely providing tools for humans to communicate and col-laborate more effectively.
However, negative emotional responses can also occur if the IS have characteristics that are very similar to those of humans (e.g., “uncanny valley”, Mori (1970)). To ensure the acceptance of AUAS and thereby create successful assistance relationships, it is necessary to better understand how humans react to anthropomorphic cues and how they affect the collaboration with AUAS (Pfeuffer et al. 2018). Earlier research in the field of IS focused on technical implementation of anthropomorphic cues, such as designing the appearance and movements of robots (Duffy 2003; Walters et al. 2008) and virtual avatars. Researchers investigated the interaction between AUAS and humans, but results are often limited to supporting functions. Research to date has hardly addressed the impact of anthropomorphic cues of AUAS on the interaction in which AUAS act as intelligent social actors collaborating with human beings. It is unclear whether the positive effects of the use of anthropomorphic cues also leads to an improvement of decision quality made in the collaboration process. In particular, there is no integrated theory that allows to understand and to explain the dependencies between the anthropomorphic cues of AUAS and the decision quality that apply during collaboration.
Based on this background, our overarching research question is: In the context of group deci-sion-making under risk by collectives of humans and advanced user assistance systems, what are the effects of anthropomorphism on the quality of the decision, the satisfaction with the deci-sion, and the personal responsibility for the decision?