AI companions (AICs; persistent digital personas driven by large language models) are increasingly popular, yet there is limited understanding of how non-experts think about the notion of companionship with machine agents.This descriptive work builds on recent scholarship by exploring topics salient to users as they discuss AI companionship in online forums.From a corpus of curated entries (N = 3,621) across 14 subreddits, LLM-assisted analysis induced four high-level themes associated with socioemotionality, control, limitations, and imaginaries.The topical hierarchy suggests this form of machine companionship both converges with and diverges from human relations.

Social robots are being groomed for human influence, including the implicit and explicit persuasion of humans. Humanlike characteristics are understood to enhance robots’ persuasive impact; however, little is known of how perceptions of two key human capacities—mind and morality—function in robots’ persuasive potential. This experiment tests the possibility that perceived robot mind and morality will correspond with greater persuasive impact, moderated by relational trustworthiness for a moral appeal and by capacity trustworthiness for a logical appeal. Via an online survey, a humanoid robot asks participants to help it learn to overcome CAPTCHA puzzles to access important online spaces—either on grounds that it is logical or moral to do so. Based on three performance indicators and one self-report indicator of compliance, analysis indicates that (a) seeing the robot as able to perceive and act on the world selectively improves compliance, and (b) perceiving agentic capacity diminishes compliance, though capacity trustworthiness can moderate that reduction. For logical appeals, social-moral mental capacities promote compliance, moderated by capacity trustworthiness. Findings suggest that, in this compliance scenario, the accessibility of schemas and scripts for engaging robots as social-moral actors may be central to whether/how perceived mind, morality, and trust function in machine persuasion.

From keeping robots as in-home helpers to banning their presence or functions, a person’s willingness to engage in variably intimate interactions are signals of social distance: the degree of felt understanding of and intimacy with an individual or group that characterizes pre-social and social connections. To date, social distance has been examined through surrogate metrics not actually representing the construct (e.g., self-disclosure or physical proximity). To address this gap between operations and measurement, this project details a four-stage social distance scale development project, inclusive of systematic item pool-generation, candidate item ratings for laypersons thinking about social distance, testing of candidate items via scalogram and initial validity analyses, and final testing for cumulative structure and predictive validity. The final metric yields a 15-item (18, counting applications with a ‘none’ option), three-dimension scale for physical distance, relational distance, and conversational distance.

As social robots' and AI agents' roles are becoming more diverse, those machines increasingly function as sociable partners. This trend raises questions about whether social gaming gratifications known to emerge in human-human co-play may (not) also manifest in human-machine co-play. In the present study, we examined social outcomes of playing a videogame with a human partner as compared to an ostensible social robot or A.I (i.e., computer-controlled player) partner. Participants (N = 103) were randomly assigned to three experimental conditions in which they played a cooperative video game with either a human, embodied robot, or non-embodied AI. Results indicated that few statistically significant or meaningful differences existed between any of the partner types on perceived closeness with partner, relatedness need satisfaction, or entertainment outcomes. However, qualitative data suggested that human and robot partners were both seen as more sociable, while AI partners were seen as more functional.