EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-SA 4.0

A Qualitative Analysis of 54 Case Studies Finds That Financial, Infrastructure, and Skills Deficits Constrain AI Capability Across Libraries and Archives

Quincy McCrary

Evidence Based Library and Information Practice · 12.06.2026 · 21 (2) · 166-168

DOI: 10.18438/eblip31044 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Evidence Based Library and Information Practice, online erschienen am 12.06.2026. Verfasst von Quincy McCrary. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

A Review of: Pinar, A., & Cox, A. (2025). An analysis of artificial intelligence (AI) capability in libraries and archives. Cataloging & Classification Quarterly, 63(6–7), 566–599. https://doi.org/10.1080/01639374.2025.2539790 Objective – This study evaluates the AI capability of libraries and archives through a qualitative content analysis of 54 case studies (2018–2024) framed by Mikalef and Gupta’s (2021) AI Capability Model. The goal is to identify how AI is used, what barriers are encountered, and where capability gaps exist. Design – Qualitative content analysis. Setting – Libraries and archives internationally across academic, national, public, and archival contexts. Subjects – 54 published case studies selected using explicit inclusion criteria that reported concrete AI applications and challenges. Methods – The authors systematically reviewed case studies identified via LIS databases (Scopus, LISA, LISTA, and Google Scholar) and extracted details of AI tools, purposes, and barriers. They analyzed the results using the AI Capability Model. Main Results – AI applications concentrated in metadata creation/enrichment and digitization (each 28% of cases), with access enhancement (20% of cases) and collection analysis (19% of cases) also present. Frequently used tools and techniques included OCR (e.g., Tesseract), NLP/transformers (e.g., BERT/KB-BERT), computer vision, machine learning classifiers, and limited use of LLMs and chatbots. The most recurrent implementation challenges were financial constraints (70% of cases), data management issues such as metadata inconsistency and noisy OCR (69% of cases), ethical concerns including bias and decontextualization (67% of cases), technical barriers like computational limits and configuration complexity (65% of cases), integration difficulties with legacy systems and workflows (54% of cases), and resource management issues including staff shortages and inadequate organizational infrastructure (39% of cases). Conclusion – The AI Capability Model provides a useful lens for assessing readiness, but libraries and archives additionally require robust ethical governance, AI literacy across staff, not only technicians, and stronger interoperability. Sector-wide capability is limited: many institutions can pilot AI but lack infrastructure, funding, and organizational change capacity to scale responsibly. Institutions in the Global South face these challenges disproportionally, including infrastructure and resource constraints, while remaining underrepresented in the published literature.

Zitieren

APA

McCrary, Q. (2026). A Qualitative Analysis of 54 Case Studies Finds That Financial, Infrastructure, and Skills Deficits Constrain AI Capability Across Libraries and Archives. Evidence Based Library and Information Practice, 21(2), 166-168. https://doi.org/10.18438/eblip31044

Vancouver

McCrary Q. A Qualitative Analysis of 54 Case Studies Finds That Financial, Infrastructure, and Skills Deficits Constrain AI Capability Across Libraries and Archives. Evidence Based Library and Information Practice. 2026;21(2):166-168. doi:10.18438/eblip31044

BibTeX
@article{McCrary2026,
  author = {McCrary, Quincy},
  title = {A Qualitative Analysis of 54 Case Studies Finds That Financial, Infrastructure, and Skills Deficits Constrain AI Capability Across Libraries and Archives},
  journal = {Evidence Based Library and Information Practice},
  year = {2026},
  volume = {21},
  number = {2},
  pages = {166--168},
  doi = {10.18438/eblip31044}
}
Kostenlos starten →