The next speakers in this session at the IAMCR 2025 conference in Singapore are Jiawei Dai and Mengyao Liu, whose interest is in human perceptions of AI. These are often shaped in a socialised manner: different cognitive frameworks will affect how people perceive AI, and what relationships they form with AI technologies. Heuristic processing provides one explanation for this: this focusses on easily noticeable and understandable cues.
Users might perceive AI chatbots as human-like, and engage with them accordingly; in fact, there are lower social expectations about how AIs might react to human users, and humans might therefore engage in greater self-disclosure than they would in social contexts with other humans. This can also be beneficial: AI technologies might foster a more inclusive atmosphere, and create greater opportunities for participation by marginalised groups, than human-to-human social contexts.
This project engaged in interviews with some 26 participants in Hong Kong, exploring their AI relationships. They experienced AI chatbots as objective but sycophantic interlocutors, regarding it as more than a machine; they also felt that they did not treat the AI chatbots as a machine, but projected a more idealised relationship on the machine. AI facilitated a reduction of undersociality: it served as a non-judgmental platform where they felt more able to express themselves freely without concern.
AI also broke the asymmetric learning circle: normally, those with more social knowledge will gain more social knowledge in future, but here those with less social knowledge learnt from the AI about how to engage more efficaciously in human-to-human engagement.
