Designing AI-Powered Reference Literacy Labs for Library Education

Thesis Abstract

The rapid digitization of information services and the increasing complexity of user information needs underscore a widening gap between traditional reference instruction and contemporary learner expectations, necessitating the integration of AI-powered labs to enhance reference literacy in library education. This study investigates how AI-driven reference literacy labs can augment the competencies of pre-service librarians and improve user outcomes in university libraries. The aim is to design, implement, and evaluate a scalable AI-enabled lab environment that supports evidence-based information retrieval, source evaluation, and ethical information use. Specific objectives are (1) to identify core reference literacy competencies in AI-assisted contexts; (2) to develop a modular AI-powered lab curriculum aligned with library accreditation standards; (3) to implement a pilot lab across three library education programs to assess feasibility and acceptability; (4) to evaluate learning gains in reference literacy using a mixed-methods framework; and (5) to formulate scalable guidelines for integration into library curricula. A convergent parallel mixed-methods design will be employed. The study will be conducted at three accredited university libraries hosting Master of Library and Information Science programs, with a total population of 180 student-teachers and 12 supervising faculty members. A stratified random sample of 120 students will be invited to participate, with 96 completing quantitative assessments and 24 participating in qualitative interviews. Data collection will involve (i) a standardized reference literacy assessment consisting of performance tasks and multiple-choice items adapted from the ACRL Information Literacy Framework, (ii) an AI-lab usage log capturing interaction patterns, task completion times, and accuracy rates, and (iii) semi-structured interviews and focus groups exploring perceived usefulness, pedagogical impact, and ethical considerations. Instrument validity will be established through content validity panels with five library education experts and a pilot test with 20 students. Reliability will be assessed using Cronbach’s alpha for the knowledge tests (target ? ? .80) and inter-rater reliability for scoring performance tasks (? ? .75). Data will be analyzed using a combination of quantitative and qualitative methods. Descriptive statistics will profile participant characteristics and baseline competencies. Inferential analyses will include multiple regression to examine predictors of learning gains, ANCOVA controlling for prior knowledge, and paired-sample t-tests to compare pre- and post-intervention scores. A thematic analysis will be conducted on interview and focus group transcripts, guided by Braun and Clarke’s approach, to identify emergent themes related to cognition, collaboration, and ethics in AI-mediated reference literacy. The study will also apply a modified technology acceptance model (TAM) to interpret user acceptance, complemented by Rogers’ Diffusion of Innovations theory to assess adoption potential within curricula. A conceptual model will be developed to illustrate the relationships among AI lab design features, learner engagement, and outcomes in reference literacy. Expected findings include significant improvements in reference literacy competencies among participants exposed to the AI-powered lab, evidenced by higher post-test scores (expected effect size d ? 0.50) and faster task completion without compromising accuracy. Qualitative data are anticipated to reveal enhanced critical evaluation of sources, improved search strategy formulation, and heightened awareness of ethical and bias-related considerations in AI-driven recommendations. The study is expected to uncover variability in uptake based on prior digital literacy, with greater gains among students who engage in reflective practice facilitated by embedded feedback mechanisms. The study contributes to knowledge by providing a theoretically informed, scalable framework for AI-enabled reference literacy education, including curriculum design guidelines, assessment rubrics, ethical safeguards, and implementation protocols for library school programs. It extends existing literature on AI in information literacy by combining pedagogical design with empirical evaluation of learning outcomes in real-world library education settings. The main conclusion is that AI-powered reference literacy labs can substantially enhance librarian-in-training competencies when integrated with structured pedagogy, rigorous assessment, and ongoing professional development for instructors. Recommendations include adopting modular AI lab components across LIS curricula, investing in instructor training on AI pedagogy and data ethics, establishing longitudinal studies to assess long-term retention and professional practice impact, and developing shared repositories of assessment tools and case-based scenarios to support cross-institutional implementation.