The communicative abilities of chatbots based on neural language models enable increasingly human-like interactions, potentially fostering the development of emotional attachment-particularly among individuals with specific psychological traits for whom virtual interactions are perceived as more accessible, predictable, and less complex than traditional interpersonal exchanges.The present study investigated the relationship between autistic traits and emotional attachment to chatbots, examining the mediating role of emotional (social anxiety and loneliness) and cognitive (anthropomorphism and theory of mind) factors.A sample of 327 young adults (aged 18-35) completed an anonymous questionnaire comprising validated measures of the variables of interest.Data analysis was conducted using structural equation modeling.Results revealed that emotional attachment to chatbots was primarily mediated by cognitive factors: anthropomorphism exhibited a positive effect, whereas reduced theory of mind capacity showed a negative effect.Emotional factors did not emerge as significant mediators.A modest direct effect of autistic traits on emotional attachment was also observed, suggesting that chatbots may serve as an accessible and meaningful relational channel for individuals with neurodivergent profiles.
Suriano et al. (Sun,) studied this question.