Abstract
Robot abuse encompasses a range of actions enacted by users against robots that resemble aggressive actions enacted by humans against each other. Research in this area faces numerous challenges due to ethical concerns and methodological limitations. Despite preliminary indications of robot abuse, most people are still not personally familiar with robots but hold collective imaginaries promoted by science fiction. Little is known about how abuse will manifest in practice, and research is in dire need of operational and theoretical definitions. As a response, this paper applies theory-driven speculative design to collect imaginaries of robot abuse with a sample of university students (N = 69). The inductive analysis revealed four behavioural patterns (Outburst, Clash, Oppression, and Rebellion) that were then operationalised in the orthogonal space defined by the Function of Aggression (reactive vs. proactive) and Robot Identity (object vs. subject). We ground the Space of Robot Abuse on the General Aggression Model and a sociotechnical definition of robots, provide a method for designing storyboards and evaluate them. We conclude by advancing a sociotechnical definition of robot abuse, which could interest researchers and designers alike.