Adoption of AI-Powered Chatbots for User Services: Challenges and Opportunities in University Libraries of Khyber Pakhtunkhwa
Abstract
Purpose: This study investigates the adoption of AI powered chatbots for user services among library professionals in HEC recognized public sector universities in Khyber Pakhtunkhwa (KP), Pakistan, examining awareness, current adoption, perceived usefulness and opportunities, implementation challenges, organizational readiness, and future adoption intention. Methodology: A cross sectional quantitative survey was conducted within a positivist research paradigm. The study population comprised 150 professional library staff across 34 public sector universities in KP. Using the Krejcie and Morgan (1970) sampling method, a sample of 108 respondents was determined. Data were collected through a structured 35 item questionnaire using a five point Likert scale, demonstrating high internal consistency (Cronbach’s α = 0.884). Data were analyzed using descriptive statistics, Pearson correlation, and multiple regression. Findings: The findings reveal a pronounced gap between awareness and institutional adoption. Awareness of AI powered chatbots was relatively high (M = 3.81), whereas current adoption remained low (M = 1.92). Perceived usefulness and opportunities were highly rated (M = 4.41), while implementation challenges were substantial (M = 4.20), particularly insufficient funding (M = 4.72), inadequate IT infrastructure (M = 4.60), and the absence of formal AI policies. Regression analysis showed that perceived usefulness (β = 0.380, p < 0.001), organizational readiness (β = 0.320, p < 0.01), and staff competency (β = 0.195, p < 0.05) significantly predicted adoption intention, explaining 57.0% of its variance (R² = 0.570). Originality: This study provides a localized empirical assessment of AI chatbot adoption in public sector university libraries in KP, Pakistan, addressing an important gap in evidence concerning conversational AI adoption in developing country higher education contexts. The findings demonstrate that limited adoption is driven less by professional unwillingness than by institutional, financial, technological, and policy constraints, highlighting the need for coordinated organizational support and governance to facilitate sustainable AI integration in academic library services.