APPLYING AI-ASSISTED CLASSIFICATION TOOLS IN ACADEMIC LIBRARY CATALOGUING: AN EVALUATION OF PRACTICAL PERFORMANCE AND LIMITATIONS IN SELECTED ACADEMIC INSTITUTIONS IN OYO STATE, NIGERIA

Abstract

This exploratory qualitative study examines how cataloguers in six private universities in Oyo State, Nigeria, perceive the practical performance and limitations of AI-assisted classification tools in academic library cataloguing. Grounded in the Technology Acceptance Model (TAM) and Knowledge Organization Theory (KOT), the study investigates four constructs: perceived usefulness, perceived ease of use, behavioural intention to adopt, and perceived challenges. Using purposive sampling, one lead cataloguer was identified at each institution; each lead cataloguer then convened a Focus Group Discussion (FGD) with colleagues directly involved in cataloguing, yielding six FGDs and a total of 27 participants. Data were analysed thematically using a combination of deductive coding, derived from the study's theoretical constructs, and inductive coding for emergent patterns. Participants generally described AI-assisted classification tools as a valuable complement to professional judgement, reporting perceived gains in metadata accuracy, consistency, and efficiency, alongside persistent challenges related to infrastructure, occasional inaccuracy in handling local and indigenous materials, and concerns about professional identity. Because the study draws on the perceptions of a small, purposively selected sample within one Nigerian state, these findings describe how this group of cataloguers experiences and interprets AI-assisted tools; they are not evidence of the tools' objective performance and are not intended to generalise beyond the study context. The study nonetheless offers context-specific, evidence-informed recommendations for cataloguing practice, staff development, and policy in Nigerian academic libraries, and identifies priorities for larger-scale and comparative research.