Adopting ChatGPT in academic library reference services: Challenges and opportunities

Background The adoption of ChatGPT in reference services delivery among academic libraries is perceived as an innovation aimed at replacing traditional reference services. Purpose This study examines the challenges and opportunities of implementing ChatGPT in reference service delivery within academic libraries. Methods A systematic review was conducted using the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta‐Analyses) framework. Various databases were consulted, including DOAJ, EBSCOhost, Google Scholar, and Emerald. A total of 123 articles were retrieved, of which 47 (38.2%) met the selection criteria. The study is guided by the theory of diffusion of innovation. The theory provides effective frameworks for explaining the adoption and use of technology inorganisations. Results The findings revealed that ChatGPT offers several benefits when integrated into reference service delivery.These benefits include its ability to provide prompt responses to users, 24/7 accessibility, research assistance, support for information literacy, and information retrieval. The study established that, despite its potential for libraries, ChatGPT has several drawbacks, including a lack of privacy and security, the potential to provide incorrect answers to users, and inherent bias. Conclusion The study revealed that the integration of ChatGPT in reference service delivery across academic libraries will not completely replace the role of reference librarians, as they will be required to intervene and respond to users' queries should ChatGPT fail. Recommendations The study recommends that librarians acquire the necessary skills to use ChatGPT for providing reference services. They should also train users to become skilful information consumers with the ability to evaluate content generated by ChatGPT.