Reproduce “Lexical Indicators of Mind Perception in Human-AI Companionship” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jaime Banks, Jianghui Li
Researchers investigated how humans verbally express mind perception (MP) during interactions with AI companions, developing linguistic indicators from chat data and linking them to broader discussions about AI companionship. The study found that explicit mental state terms and novel interactional phrases signal MP, with a positive association between using MP language in chat and discussing AI limitations and existential topics.
Key insights
- Mind perception in direct human-AI chat is conveyed through a concentrated set of cognitive and emotional terms, alongside specific interactional expressions that reflect a relational stance.
- The commonly distinct dimensions of perceived agency and experience may become intertwined or collapse into a more holistic "Is this someone?" perception during active human-AI interaction.
- Humans engage in complex critical engagement with AI, where expressing mind-attributing language in chat can coexist with critical reflections on AI's artificiality and limitations in other discussions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.