The Governance Gap: A Systematic Review of the Adequacy of AI Governance Frameworks for Global Academic Libraries

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Academic libraries are adopting AI for discovery, cataloguing, reference, and analytics faster than governance instruments have evolved. This systematic review evaluated whether international, national, and professional AI governance frameworks adequately address academic-library needs, examined accountability structures for AI-related harms, and explored differences between high-income and low- and middle-income contexts. The protocol was registered prospectively with the Open Science Framework on 2 March 2026 (https://osf.io/mvgbf) and the review followed PRISMA 2020. Searches (January 2015 to March 2026) spanned eight bibliographic databases and six grey-literature sources. From 2,847 de-duplicated records, 412 full texts were assessed and 69 studies met inclusion (Cohen’s kappa = 0.79 title/abstract; 0.83 full text; 0.74 dimensional coding). Data were synthesized using framework analysis across ten governance dimensions and thematic synthesis across six themes. Assigned geographic focus was uneven: 43% North America, 32% Europe, 13% East and South-East Asia and Australasia, 9% Africa, and 3% global comparative; no included study took a South Asian, Latin American, Caribbean, Middle Eastern, or Central Asian library context as its primary frame. No framework addressed all ten dimensions. Where a dimension was addressed, more than nine in ten instances failed the review’s two-condition adequacy criterion, the recurring deficiencies being absent operational specificity, absent enforcement, and library-contextual silence. Library-specific provisions satisfied the criterion in 2 of the 23 frameworks analyzed (8.7%), and no study documented an instrument meeting it on intellectual freedom (0 of 31). Confidence in the five synthesized gap patterns, appraised with GRADE-CERQual, was high for two, moderate for one, and low for the two concerning vendor accountability and Global South equity. The review recommends operationally specific, internationally coordinated governance that includes Global South libraries, and procurement regulation for AI vendors serving publicly funded libraries as a near-term intervention.