Developing Locally-Grounded Ethical Artificial Intelligence Frameworks for South African Academic Libraries
DOI:
https://doi.org/10.22373/adabiya.v28i2.36354Keywords:
Artificial intelligence, South African academic libraries, Epistemic justice, Algorithmic bias, Artificial intelligence governanceAbstract
This paper explores the real-world ethical challenges and governance gaps that South African academic libraries face as they adopt Artificial Intelligence (AI). Through conversations with library directors, focus groups with frontline librarians, and analysis of information technology (IT) policy documents from six universities, the study uncovers four main problem areas: absence of local, relevant AI governance; algorithms often overlook African and Global South scholarship; staff don’t always have the training or direction they need; and librarians themselves are often sidelined, expected to use new technologies rather than shape how they are used. To address these issues, the study offers a practical Ethical AI Integration Framework (EAIF), built on five key pillars: epistemic justice, algorithmic transparency, data sovereignty, multilingual equity, and participatory stewardship. Ultimately, the research argues that South African academic libraries need to develop their own ethical approaches rooted in local values, languages, and a commitment to fair access to knowledge rather than relying on imported frameworks that may not fit their contexts. This study provides, for the first time an evidence-based, co-designed governance model for AI adoption in African libraries by shifting the narrative beyond mere description of adoption towards providing the tools for responding to the epistemic exclusion, multilingual silencing and de-skilling in an interdependent yet distinct, complex cluster of ethical, infrastructural and agentic issues neglected by existing work which has tended to foreground access and equipment over ethical implementation.
Abstrak
Makalah ini mengeksplorasi tantangan etika dan kesenjangan tata kelola di dunia nyata yang dihadapi perpustakaan akademik Afrika Selatan saat mereka mengadopsi Kecerdasan Buatan (AI). Melalui percakapan dengan direktur perpustakaan, kelompok fokus dengan pustakawan garis depan, dan analisis dokumen kebijakan teknologi informasi (TI) dari enam universitas, studi ini mengungkap empat area masalah utama: kurangnya tata kelola AI lokal yang relevan; algoritma sering mengabaikan kajian ilmiah Afrika dan Global Selatan; staf tidak selalu memiliki pelatihan atau arahan yang mereka butuhkan; dan pustakawan sendiri sering kali dikesampingkan, diharapkan untuk menggunakan teknologi baru daripada membentuk bagaimana teknologi tersebut digunakan. Untuk mengatasi masalah ini, studi ini menawarkan Kerangka Integrasi AI Etis (EAIF) yang praktis, yang dibangun di atas lima pilar utama: keadilan epistemik, transparansi algoritma, kedaulatan data, kesetaraan multibahasa, dan tata kelola partisipatif. Pada akhirnya, penelitian ini berpendapat bahwa perpustakaan akademik Afrika Selatan perlu mengembangkan pendekatan etika mereka sendiri yang berakar pada nilai-nilai lokal, bahasa, dan komitmen terhadap akses yang adil terhadap pengetahuan daripada bergantung pada kerangka kerja impor yang mungkin tidak sesuai dengan konteks mereka. Studi ini menyajikan, untuk pertama kalinya, model tata kelola berbasis bukti dan dirancang bersama untuk adopsi AI di perpustakaan Afrika dengan menggeser narasi dari sekadar deskripsi adopsi menuju penyediaan alat untuk menanggapi eksklusi epistemik, pembisuan multibahasa, dan penurunan keterampilan dalam kelompok isu etika, infrastruktur, dan agensi yang saling bergantung namun berbeda dan kompleks, yang diabaikan oleh karya-karya yang ada yang cenderung mengedepankan akses dan peralatan daripada implementasi etis.
Downloads
References
Adelani, D. I., Alabi, J., Alkhamissi, B., Ojo, A., Emezue, C., Nasution, R., ... & Ruder, S. (2023). MasakhaNER 2.0: Africa-centric transfer learning for named entity recognition. Transactions of the Association for Computational Linguistics, 11, 21–37. https://doi.org/10.1162/tacl_a_00536
American Library Association. (2021). ALA statement on artificial intelligence. https://www.ala.org/tools/ethics/artificial-intelligence
Anderson, J., & Rainie, L. (2022). AI and the future of humans: Expert perspectives on governance. Pew Research Centre.
Andrade, M., & Alves, P. (2022). Participatory action research in library settings: A methodological review. Journal of Librarianship and Information Science, 54(3), 412–429. https://doi.org/10.1177/09610006211015022
Bawden, D., & Robinson, L. (2022). Introduction to information science (2nd ed.). Facet Publishing.
Benjamin, R. (2020). Race after technology: Abolitionist tools for the new Jim Code. Polity Press.
Braun, V., & Clarke, V. (2022). Thematic analysis: A practical guide. SAGE Publications.
Breed, M. (2023). Library systems report 2023: Advancing AI integration. American Libraries, 54(5), 22–35.
Brock, A. (2020). Critical technocultural discourse analysis. New Media & Society, 22(6), 1012–1030. https://doi.org/10.1177/1461444820914736
Brown, C., & Czerniewicz, L. (2020). Debunking the 'digital native': Beyond digital apartheid, towards digital democracy. Journal of Computer Assisted Learning, 36(5), 583–594. https://doi.org/10.1111/jcal.12476
Buckingham, D. (2020). A critical engagement with digital literacy. Journal of Media Literacy Education, 12(2), 114–119.
Choi, W., An, S., & Han, G. (2022). AI-powered reference services: User satisfaction and interaction quality. College & Research Libraries, 83(4), 512–530. https://doi.org/10.5860/crl.83.4.512
Chukwuemeka, A., & Ugwuanyi, C. (2023). Artificial intelligence and the changing roles of academic librarians in Africa. African Journal of Library, Archives and Information Science, 33(1), 45–61.
Couldry, N., & Mejias, U. A. (2021). Data colonialism: Rethinking big data's relation to the contemporary subject. Television & New Media, 20(4), 336–349. https://doi.org/10.1177/1527476418796632
Creswell, J. W., & Poth, C. N. (2022). Qualitative inquiry and research design: Choosing among five approaches (5th ed.). SAGE Publications.
Czerniewicz, L., Agherdien, N., Badenhorst, J., Bhatt, D., Chili, M., De Villiers, M., ... & Wiens, K. (2021). A wake-up call: COVID-19 and its impact on the core business of higher education institutions in South Africa. Studies in Higher Education, 46(1), 1–16. https://doi.org/10.1080/03075079.2020.1813351
Department of Higher Education and Training. (2022). Digital transformation in higher education: A strategic framework 2022–2027. Government Printer.
Department of Science and Innovation. (2022). South Africa's national artificial intelligence policy framework. Government Printer.
Dlamini, N., & Ngoepe, M. (2021). Digital transformation in South African public institutions: Governance challenges and opportunities. South African Journal of Information Management, 23(1), 1–11. https://doi.org/10.4102/sajim.v23i1.1312
Fricker, M. (2007). Epistemic injustice: Power and the ethics of knowing. Oxford University Press.
Gorichanaz, T. (2023). The librarian as ethical AI steward: Professional identity in the age of algorithmic systems. Library Quarterly, 93(2), 101–118. https://doi.org/10.1086/724021
Helgesson, G., & Eriksson, S. (2021). Plagiarism in research and the inequitable treatment of scholars from the Global South. Research Ethics, 17(1), 5–18. https://doi.org/10.1177/1747016120954412
International Federation of Library Associations and Institutions. (2022). IFLA statement on libraries and artificial intelligence. https://www.ifla.org/publications/node/93554
Islam, M. A., & Tsuji, K. (2021). AI-driven reference services: A review of implementations and outcomes. Journal of the American Society for Information Science and Technology, 72(8), 1093–1107. https://doi.org/10.1002/asi.24453
Library and Information Association of South Africa. (2023). State of academic libraries in South Africa: Annual survey report 2023.
LIASA.
Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. SAGE Publications.
Luthmann, A., & Park, J. (2022). Automated metadata enrichment using natural language processing: Challenges for multilingual collections. Cataloguing & Classification Quarterly, 60(5), 348–366. https://doi.org/10.1080/01639374.2022.2056128
Masakhane. (n.d.). Masakhane: African languages in AI. https://www.masakhane.io/
Mohamed, S., Png, M. T., & Isaac, W. (2020). Decolonial AI: Decolonial theory as sociotechnical foresight in artificial intelligence. Philosophy & Technology, 33(4), 659–684. https://doi.org/10.1007/s13347-020-00405-8
Molepo, M., & Fourie, I. (2020). Digital divide, digital literacy and academic libraries in South Africa. South African Journal of Libraries and Information Science, 86(2), 1–14. https://doi.org/10.7553/86-2-1810
Munn, L. (2022). The uselessness of AI ethics. AI and Society, 37(4), 1–6. https://doi.org/10.1007/s00146-021-01180-8
Ndlovu-Gatsheni, S. J. (2021). Decolonisation, development and knowledge in Africa: Turning over a new leaf. Routledge.
Ndovu, T., & Mutula, S. (2021). AI governance in sub-Saharan African public institutions: A systematic review. Information Development, 37(4), 617–635. https://doi.org/10.1177/02666669211020428
Noble, S. U. (2021). Algorithms of oppression: How search engines reinforce racism (2nd ed.). New York University Press.
Obermeyer, Z., Powers, B., Vogeli, C., & Mullainathan, S. (2022). Dissecting racial bias in an algorithm used to manage the health of populations. Science, 366(6464), 447–453. https://doi.org/10.1126/science.aax2342
OCLC. (2023). Artificial intelligence in libraries: A global benchmark study. OCLC Research. https://www.oclc.org/research/publications/2023/oclc-ai-libraries-global.html
Okonkwo, C., & Ezema, I. J. (2022). Challenges and prospects of AI adoption in African academic libraries. Library Hi Tech, 40(3), 718–735. https://doi.org/10.1108/LHT-05-2021-0176
Olteanu, A., Castillo, C., Diaz, F., & Kıcıman, E. (2020). Social data: Biases, methodological pitfalls, and ethical boundaries. Frontiers in Big Data, 2, 13. https://doi.org/10.3389/fdata.2019.00013
Park, J., & Lu, C. (2021). Machine learning for subject cataloguing: A review of approaches and implications for metadata professionals.
Library Resources & Technical Services, 65(1), 34–52. https://doi.org/10.5860/lrts.65n1.34
Participatory Research in Asia. (2021). Participatory action research: Principles and practice (3rd ed.). PRIA.
Pawley, A. L. (2022). Unequal legacies: Race and multiculturalism in the LIS curriculum. Library Quarterly, 92(3), 269–291. https://doi.org/10.1086/719978
Raber, D., & Budd, J. M. (2022). Information as sign: Semiotics and information science. Journal of Documentation, 78(1), 1–18. https://doi.org/10.1108/JD-04-2020-0054
Republic of South Africa. (2020). Presidential Commission on the Fourth Industrial Revolution: Final report. The Presidency.
Sadeh, T. (2022). The evolution of library discovery systems. Information Technology and Libraries, 41(2), 1–22. https://doi.org/10.6017/ital.v41i2.14219
Sewchurran, K., & Gregoriou, P. (2022). Ethical technology adoption in South African higher education: A critical discourse analysis. South African Journal of Higher Education, 36(3), 212–231. https://doi.org/10.20853/36-3-4852
South African Institute of Race Relations. (2023). South Africa survey 2023. SAIRR.
Tammaro, A. M., Dick, A., & Salarelli, A. (2021). Artificial intelligence and libraries: Toward a future agenda for research and practice. JLIS.it, 12(3), 1–12. https://doi.org/10.4403/jlis.it-12740
Tladi, L. (2021). Privacy rights in the digital age: A South African perspective. Potchefstroom Electronic Law Journal, 24(1), 1–27. https://doi.org/10.17159/1727-3781/2021/v24i0a8784
UNESCO. (2021). Recommendation on the ethics of artificial intelligence. United Nations Educational, Scientific and Cultural Organisation. https://doi.org/10.54676/GPPC3851
Universities South Africa. (2024). Higher education sector statistics 2023. USAf.
Vakkari, P. (2023). Searching and information seeking in digital environments: An overview. Information Processing & Management, 60(1), 103153. https://doi.org/10.1016/j.ipm.2022.103153
Vieira, D. V., & Garcez, P. M. (2021). Machine learning for collection development in academic libraries: A predictive model for acquisition decisions. Library & Information Science Research, 43(4), 101121. https://doi.org/10.1016/j.lisr.2021.101121
World Bank. (2024). Poverty and inequality platform: South Africa country profile 2023. World Bank Group. https://pip.worldbank.org/country-profiles/ZAF
Zhang, Y., & Arlitsch, K. (2022). Predictive analytics for library collection management: Methodologies and outcomes. Library Collections, Acquisitions, & Technical Services, 45(1), 1–14. https://doi.org/10.1080/14649055.2021.1989762
Zou, J., & Schiebinger, L. (2021). AI can be sexist and racist—it's time to make it fair. Nature, 559(7714), 324–326. https://doi.org/10.1038/d41586-018-05707-8
Downloads
Published
License
Copyright (c) 2026 Dickson Mdhlalose (Author)

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Procedure Proposed for Public Access Journals The authors of Jurnal Adabiya Journal adhere to the following terms:
a. Authors retain copyright and grant the journal the right of first publication, with the work licensed under a Creative Commons Attribution-ShareAlike 4.0 International License that permits others to share the work with attribution to the authorship and first publication in this journal.
b. Authors may enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (such as posting it to an institutional repository or publishing it in a book), with attribution to the journal's initial publication.
c. Authors are permitted and encouraged to submit their work online (e.g., in institutional repositories or on their website) before and during the submission process, as this can lead to fruitful exchanges and increased citations of published work (See The Effect of Open Access).
d. Jurnal Adabiya Journal publishes, distributes, uses, and repurposes scholarly works under the CC-BY SA license or an equivalent license.
