ARTIFICIAL INTELLIGENCE IN LIBRARY SYSTEMS IN KENYA: A SCOPING REVIEW WITH GLOBAL AND AFRICAN CONTEXTUAL INSIGHTS

BACKGROUND AND RATIONALE Artificial Intelligence (AI) technologies, including machine learning systems, natural language processing tools, generative models such as GPT variants, robotic process automation, and predictive analytics, are increasingly shaping how information is created, classified, retrieved, and interpreted (Florida, 2024; Makinde & Oladokun, 2021.). In many sectors, these tools have moved from experimental pilots to operational systems (Orubebe et al., 2025). Libraries, as longstanding public knowledge institutions, are not immune to these shifts. Whether public, academic, school-based, or special collections, libraries continue to support teaching, research, cultural memory, and community access to information (Sousa, 2025). The question, however, is not simply whether AI is present, but how it is being integrated, and under what conditions (Boateng, 2025). In high-income countries, documented applications are no longer limited to experimental prototypes. Automated metadata generation through platforms such as OCLC's AI-supported cataloging tools, chatbot-based user services (e.g., LibraryH3lp deployments), AI-assisted plagiarism detection embedded in systems like Turnitin, and predictive analytics for collection management have been reported in both practice literature and peer-reviewed studies (Florida, 2024; Guo et al., 2025). These cases often describe gains in processing time, user responsiveness, or collection forecasting accuracy (Orubebe et al., 2025). That said, reported benefits are not always accompanied by detailed cost analyses or long-term evaluations. In some instances, claims of efficiency appear to rely on short-term institutional reporting rather than independent assessment (Boateng, 2025). Continental and national digital policy frameworks, such as the African Union's Digital Transformation Strategy (2020–2030) and Kenya's emerging National AI Strategy (2025–2030), signal political interest in AI adoption (Orubebe et al., 2025). Kenya's broader digital infrastructure ambitions, including the Digital Superhighway initiative under Vision 2030, further underscore this trajectory (Makinde & Oladokun, 2021). Yet, within African library systems, implementation conditions may differ considerably from those described in North American or European literature (Makinde & Oladokun, 2021). Available accounts suggest infrastructural instability in some regions (including inconsistent electricity supply or uneven broadband connectivity), uneven professional development opportunities in AI-related skills, and limited institutional policy guidance on algorithmic governance (Orubebe et al., 2025; Makinde & Oladokun, 2021). In Kenya specifically, digitization initiatives led by the Kenya National Library Service (KNLS), university consortia supported through KENET, and small-scale experimentation with AI tools are emerging (Makinde & Oladokun, 2021). Still, there appears to be no consolidated synthesis mapping the scale, readiness levels, or governance maturity of these efforts (Orubebe et al., 2025). In informal consultations during the development of this protocol, librarians noted enthusiasm about AI tools, particularly generative AI for user reference queries, alongside concerns about licensing costs, data sovereignty, and staff training burdens (Boateng, 2025). These tensions, while anecdotal, underscore the need for a systematic mapping of available evidence (Guo et al., 2025).