Abstract

Anthropomorphic agents with increasing levels of autonomy are being used in a growing number of applications. This is especially evident in games where characters are designed with human likeness both in appearance and behaviour, with a level of autonomy that allows them to surprise and engage the player. However, with these autonomous system there is the possibility that non-intended behaviours may emerge, exposing the user to potentially ethically questionable encounters. In this position paper we argue for further protections against such glitches through the implementation of artificial ethics-based behavioural safeguards. We begin by outlining the background and specific challenges of this emerging field, before proposing a direction for future research. We conclude with a call to action, arguing that significant cross-disciplinary research, and engagement from the HCI community is required in this area.