Presented at the Trustworthy Autonomous Systems Symposium 2024, September 16–18, 2024, Austin, TX, USA. This paper is an output from Horizon’s FAILSAFE project (a Consumer Products Campaign project).
Harriet Cameron, Simon Castle-Green, Muhammad Chughtai, Liz Dowthwaite, Ayse Kucukyilmaz, Horia Maior, Victor Ngo, Eike Schneiders, Bernd Stahl
This papers addresses that domestic robots are fast becoming an integrated part of daily life. In anticipation of increased uptake of robotic assistants in the home, researchers and designers must investigate what makes domestic robotic interventions trustworthy or untrustworthy as a matter of urgency. This paper explores the concept of failure in domestic robotics, using the case of a dishwasher robot, and its impact on trustworthiness. It asks what constitutes trust, what constitutes failure, and what are the impacts failure may have on service providers, users, and the robot itself. We present the findings from four workshops with robotics experts and potential end users. We show that failure is simultaneously complex and predictable and re-evaluate existing taxonomies of failure, applying them to the domestic sphere, thereby highlighting social and corporate facets of failure that are not currently represented. We also provide a new taxonomy of failure outcomes to highlight how failures can breach trust, and what effects that breach may have.