De-instrumentalizing HCI

social psychology, rapport formation, and interactions with artificial social agents

Ritwik Banerji

pp. 43-66

in: Michael Filimowicz, Veronika Tzankova (eds), New directions in third wave human-computer interaction I, Berlin, Springer, 2018

Abstract

Decisions in designing artificial social interactants to reproduce culturally-specific forms of human sociality evince a range of conceptions of the norms and cognitive processes involved in the human social interactions themselves. Regarding the use of machine learning (ML) in such systems, decisions whether or not to use this approach implicitly presents questions on the nature of the interpersonal adaptation that takes place and indicate a range of conceptions of the values which structure these interactions. In the design of virtual performers of musical free improvisation, several designers assume that the experience of equal partnership between improvisers can only be afforded through deployment of ML in such systems. By contrast, tests of agents not based in ML reveal that human beings experience illusions of "adaptation" in interactions with systems which lack any adaptive capacity. Such results suggest that HCI research with artificial social interactants may be used to raise new questions about the nature of human interaction and interpersonal adaptation in the formation of relationships over time.