AI-Enabled Library Services and User Satisfaction: Investigating Service Quality, Accessibility, and Ethical Challenges

Abstract

Artificial intelligence (AI) has moved from an experimental curiosity to an operational reality in library and information centres, reshaping how catalogues are searched, how reference queries are answered, and how collections are recommended to readers. This study investigates the relationship between AI-enabled library services and user satisfaction, with particular attention to three dimensions that determine whether such systems succeed or fail in practice: service quality, accessibility, and ethical trustworthiness. Adopting a descriptive survey design, data were collected from users of academic and public libraries through a structured questionnaire built on an adapted SERVQUAL framework supplemented by accessibility and ethics constructs derived from recent literature. The instrument measured user awareness of AI-enabled services, frequency of use, perceived service quality across reliability, responsiveness, assurance, empathy and tangibility, perceived accessibility across infrastructural and competency dimensions, and perceived ethical risk across privacy, transparency and algorithmic bias. Five hypotheses were formulated and tested using correlation analysis, independent samples t-tests, one-way analysis of variance, chi-square tests of independence, and multiple linear regression. The findings indicate that perceived service quality is the strongest positive determinant of user satisfaction, that accessibility barriers significantly moderate the benefit users derive from AI-enabled services, and that ethical concerns—particularly around data privacy and the opacity of recommendation systems—exert a measurable negative influence on user trust and, through it, on satisfaction. Differences between academic and public library users emerged in awareness and intensity of use rather than in overall satisfaction, suggesting that the determinants of a positive AI service experience are broadly consistent across library types. The study concludes that AI adoption in libraries yields satisfaction dividends only when deployment is accompanied by deliberate investment in infrastructure, user training, transparent data governance, and the retention of human mediation as an accessible alternative. Recommendations are offered for library administrators, system vendors, and professional associations.