Lexical Analysis of 55,968 Online Reviews Reveals Key Factors in Human-AI Interactions

A new preprint on arXiv (ID 2511.13480) presents a lexical analysis of 55,968 online reviews to understand the dynamics between humans and AI systems. The study, categorized under Computer Science > Human-Computer Interaction, addresses a gap in existing research on human-AI interaction, which has previously focused on user perceptions and ethical considerations but not on the specific concerns and challenges users face. By applying a lexical approach to reviews from three unnamed platforms (indicated as this http URL), the researchers conducted factor analysis to identify key factors influencing these interactions. Initial results from the factor analysis reveal these factors, and the study plans to use content analysis for deeper insights. The findings aim to contribute to the development of more user-centric AI systems and to inform future AI technology and user experience improvements. The paper is part of the arXiv submission history and includes references and citations via Semantic Scholar, as well as tools for bibliographic citation and code/data sharing. It also mentions arXivLabs, a framework for collaborative experimental projects, which adheres to values of openness, community, excellence, and user data privacy.