Mind the gap : exploring AI expectations of librarians and patrons at the University of Malta in the Gen-AI era

Abstract

Purpose: The emergence of Generative Artificial Intelligence (GenAI) in late 2022 precipitated a new academic reality, fundamentally altering information behaviors in higher education. While the COVID-19 pandemic accelerated the shift toward remote digital services, the GenAI era challenges the core function of the academic library: the mediation of information discovery and verification. Current discourse within Library and Information Science (LIS) often treats librarian readiness and student adoption as separate phenomena. However, anecdotal evidence and emerging literature suggest a growing disconnect, an expectation gap, between the services libraries are preparing to offer (often focused on backend infrastructure and copyright governance) and the services students actually require (often focused on workflow efficiency, citation assistance, and grade preservation). The primary purpose of this research is to bridge this provider-only perspective by quantifying the divergence between Library Professionals (Providers) and Library Patrons (Users) at the University of Malta Library (UML). By juxtaposing professional assumptions against user realities, the study aims to: Identify the specific areas where student usage of AI outpaces staff awareness (The Usage Gap). Determine if the service interventions proposed by staff align with the immediate "pain points" of the student body (The Service Gap). Propose a "Service Alignment Model" to help the UML prioritize AI initiatives that yield the highest value to the user community, moving beyond reactive policy-making toward proactive, user-centered service design.

Design/methodology/approach. This study employs a quantitative, cross-sectional, dual-perspective research design method. Rather than relying on disparate datasets collected at different times, this research utilizes a mirrored survey methodology to capture the concurrent viewpoints of library professionals (Providers) and library patrons (Users). Two structured questionnaires will be distributed simultaneously over a four-week period to two respondent cohorts: Cohort A (Providers): targeting UML staff members, measuring their readiness, willingness to train, and perceptions of student behavior. Cohort B (Users): targeting registered UM patrons (undergraduate and postgraduate students, academics and administrative staff), measuring their actual usage patterns, service expectations, and ethical priorities. The instruments are designed with identical underlying constructs to allow for a direct paired gap analysis. Key measurement areas include: Usage Constructs: Comparing staff estimates of AI adoption against patron self-reported usage frequency, specifically probing for concealment behaviors (hiding AI use due to fear of penalization) and discovery substitution (using LLMs instead of the library catalog). Service Constructs: Mapping patrons' demands to emerging Intelligent Library staff roles (e.g., AI literacy workshops, verification clinics) and comparing these demands against staff willingness and readiness to provide these new services. Ethical Constructs: Ranking ethical concerns to measure the divergence between provider priorities (data privacy/governance) and user priorities (grades/accuracy).

Theoretical base: This research is theoretically grounded in Customer Knowledge Management (CKM) and Service Quality (Gap Analysis) Theories. CKM logic argues that in dynamic information environments, service innovation must be driven by the systematic collection of external user knowledge rather than internal process optimization. In the context of the Intelligent Library, libraries risk obsolescence if they focus only on optimizing traditional workflows (e.g., OPAC searching) while users migrate to novel workflows (e.g., conversational approaches to searching via ChatGPT and AI search engines). Furthermore, the study draws on the Service Quality Gap Model, specifically the knowledge gap, the difference between what service providers believe customers expect and what customers actually expect. By quantifying this gap, the research moves beyond abstract discussions of AI readiness to a concrete measurement of Service Alignment. This framework is particularly relevant in the post-pandemic context, where user expectations for seamless, remote, and smart services have intensified.

Expected results: As the survey is forthcoming, results are hypothesized based on a synthesis of preliminary staff data gathered in a 2025 survey and broader global literature on student AI adoption. We anticipate three primary divergence patterns: A Usage Gap: It is expected that patron self-reported usage will significantly exceed staff estimates, particularly regarding shadow tasks. We hypothesize that a significant portion of patrons are already using AI for discovery and synthesis, bypassing the library catalog, while staff may perceive AI usage as limited to brainstorming and text generation. We also anticipate high rates of concealment, where patrons use AI but hide it due to lack of clear policy, a behavior likely unknown to staff. A Service Gap: We expect to find that staff are over-investing in theoretical readiness (e.g., willingness to learn about backend AI tools) while under-serving immediate practical needs. Specifically, we hypothesize high patron demand for verification clinics (help checking if AI citations are real) and prompt engineering for Research, which may not currently be on the library’s training roadmap. An Ethical Gap: We anticipate a sharp divergence in ethical orientation. Staff are expected to prioritize data privacy and copyright (governance concerns), consistent with professional library ethics. In contrast, patrons are expected to prioritize accuracy (hallucination risks) and grade validity (outcome concerns). This gap suggests that current library warnings about privacy may be ignored by patrons who are more focused on getting the answer right.

Originality/value: The originality of this research lies in its relational approach. While existing studies often examine librarian attitudes or student applications in isolation, this study treats them as a single service ecosystem. By using the mirrored survey methodology, it provides empirical evidence of the relationship between provider assumptions and user realities.

Practical implications: The primary practical output of this research will be a Service Alignment Model (SAM) for the UML. This framework will move the library from a reactive stance towards a proactive strategy by: Re-prioritizing Training: Shifting user training from the staff perceived need for general AI awareness training to specific competencies that patrons require, such as AI citation verification. Policy Communication: Moving from prohibitive policies (bans) to enablement policies that address patrons fears of accidental plagiarism. Strategic Positioning: Validating the library’s new value proposition not as a search engine competitor (competing on speed) but as a human-in-the-loop verifier (competing on trust). Ultimately, the study argues that closing the expectation gap is not just about adopting new technology; it is about ensuring the library remains relevant in the cognitive workflow of the modern student.