Purpose This study aims to critically examine the adoption of agentic artificial intelligence (AI) in academic libraries, focusing on use cases, the skills required, the ethical issues academic libraries should consider and future directions. Design/methodology/approach This study adopts a conceptual approach, drawing on prior literature and proposing the future adoption of agentic AI in academic libraries and its implications. Findings The study findings indicate that agentic AI could be adopted in academic libraries for event scheduling automation, weeding, serving as a reference and research agent and reducing human error. The researchers proposed that academic librarians need to reskill and upskill (agentic AI literacy, legal and policy literacy, AI-informed pedagogical skills, technical and data skills) to remain relevant in the current agentic AI era, while accounting for ethical considerations, including accountability, privacy, user agency and equity. Originality/value To the authors’ knowledge, this study is among the earliest contributions to agentic AI in academic libraries. This study contributes to the agentic AI and academic libraries by considering adoption and proposed use cases, skills, ethical issues and future directions for academic libraries. This study offers a conceptual explanation for educators, librarians and policymakers seeking to adopt agentic AI in academic libraries.
Ayinde et al. (Wed,) studied this question.