Chatbot anthropomorphism might not be the design for all: examining responses to anthropomorphized chatbots by autistic individuals

This research delves into the realm of chatbot anthropomorphism and its potential negative implications, with a specific focus on the preferences of individuals with autism. While existing literature often highlights the positive effects of anthropomorphism and attributes them to the sense of social presence it elicits, this article challenges that notion. The authors argue that these favorable effects seem to be confined to individuals who anticipate and prefer conventional interpersonal interactions embedded with rich social norms and skills. Through two experimental studies, this article reveals a counterintuitive finding: individuals with autism exhibit a preference for low-anthropomorphic chatbot agents over high-anthropomorphic ones. The primary contribution of this research lies in unveiling an adverse consequence within contemporary chatbot interface design. It signals a need for a shift toward a more inclusive approach to chatbot interface design—one that considers the cognitive and social requisites of marginalized users.