This study explores how adults with intellectual disabilities anthropomorphise social robots and construct their identities during interactions. We conducted a qualitative analysis of video data from two previous studies involving adults interacting with two robots, Pepper (a humanoid) and Cozmo (a small, toy-like robot). Using inductive thematic analysis, we examined how anthropomorphic design influenced participants’ perceptions, expectations, and social behaviours. Our findings reveal that participants readily attributed human-like qualities and identities (such as friend, companion, or learning facilitator) to the robots, shaping their interactions accordingly. Pepper’s human-like appearance raised user expectations, occasionally leading to frustration when those expectations were unmet. Conversely, Cozmo’s playful, non-humanoid design set lower expectations. Reciprocity and mirroring emerged as critical to rapport-building, while gender attribution and social roles were dynamically constructed by participants. We discuss practical implications for inclusive robot design, emphasizing the alignment of appearance with capability, clear reciprocal signals, and adaptable interaction modalities. This research advances our understanding of robo-identity, highlighting its fluidly constructed nature and offering insights into designing robots that meaningfully engage adults with intellectual disabilities.
Balasuriya et al. (Fri,) studied this question.