Artificial intelligence skills for library reference services among librarians: a systematic review

Artificial intelligence (AI) technologies comprise chatbots, recommender engines, and virtual reference systems, among others. The specific AI skills librarians need to possess to deploy these tools effectively remain poorly assessed. The study examines the AI skills needed by librarians to properly leverage these tools for their service delivery requirements. A systematic literature review guided by the PRISMA 2020 guidelines was conducted. Peer-reviewed journal articles and conference papers published between 2016 and 2026 were retrieved from Scopus, Web of Science, and two others. Studies were deliberately sourced from six world regions to maximize global representativeness. In total, 334 records were identified, and 31 studies met the inclusion criteria and were analyzed using narrative synthesis. Librarians demonstrated increasing awareness of AI-powered reference tools but exhibited limited practical competency in deploying, evaluating, and troubleshooting them. Existing strengths in information literacy, communication, and subject expertise provide a foundation for AI-mediated reference services, yet deficiencies persist in data literacy, conversational AI design, and algorithm bias recognition. Other challenges include outdated LIS curriculum, inadequate professional development, and weak institutional support, among others, hindering AI skill acquisition across most regions. This study proposes a five-component AI Reference Competency Framework specifically on AI skills for library reference services.