Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
Thesis (MPhil)--Stellenbosch University, 2026.
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
Stellenbosch : Stellenbosch University
2026
|
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867613926796034048 |
|---|---|
| access_status_str | Open Access |
| author | Kalaba, Chiyesu Joseph |
| author2 | De Villiers-Botha, Tanya |
| author_browse | De Villiers-Botha, Tanya Kalaba, Chiyesu Joseph |
| author_facet | De Villiers-Botha, Tanya Kalaba, Chiyesu Joseph |
| author_sort | Kalaba, Chiyesu Joseph |
| collection | Thesis |
| dc_rights_str_mv | Stellenbosch University |
| description | Thesis (MPhil)--Stellenbosch University, 2026. |
| format | Thesis |
| id | oai:scholar.sun.ac.za:10019.1/136148 |
| institution | Stellenbosch University (South Africa) |
| language | English |
| last_indexed | 2026-06-10T12:43:54.909Z |
| license_str | Other — see source repository |
| provenance_str_mv | Harvested via OAI-PMH from SUNScholar — Stellenbosch University Repository |
| publishDate | 2026 |
| publishDateRange | 2026 |
| publishDateSort | 2026 |
| publisher | Stellenbosch : Stellenbosch University |
| publisherStr | Stellenbosch : Stellenbosch University |
| record_format | dspace |
| source_str | SUNScholar — Stellenbosch University Repository |
| spelling | oai:scholar.sun.ac.za:10019.1/136148 AI Ethics in Africa: Postcolonial Challenges and Ethical Alternatives Kalaba, Chiyesu Joseph De Villiers-Botha, Tanya Stellenbosch University. Faculty of Arts and Social Sciences. Dept. of Philosophy. Thesis (MPhil)--Stellenbosch University, 2026. Kalaba, C. J. 2026. AI Ethics in Africa: Postcolonial Challenges and Ethical Alternatives. Unpublished masters thesis. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/79d32597-8aaa-4fe5-a4da-69cedb6d9ded This thesis develops a normative framework for artificial intelligence (AI) governance that is both contextually legitimate for Sub-Saharan Africa and interoperable with global regimes. It addresses a common dilemma: policy makers either borrow comprehensive but institutionally demanding models from the Global North or publish aspirational strategies with limited enforcement depth. The study proposes an Adapted AI Governance Framework that retains the procedural strengths of the European Union’s approach while embedding ubuntu’s relational ethics, thereby aligning enforceability with communal legitimacy. The work is conceptual and analytical. It uses argument mapping and reflective equilibrium to assess principles, provisions, and institutional feasibility. The analysis is attentive to cross-border interoperability and to the risks of misrepresenting African traditions within global academic discourse. The framework rests on five design principles: First, relational dignity grounds personhood in community and requires duties that protect social bonds. Second, an expanded account of harm captures relational harms to trust, participation, and cohesion, in addition to individual harms. Third, restorative accountability complements deterrence by requiring repair, explanation, and learning following adverse outcomes. Fourth, participatory oversight assigns defined roles to affected communities, civil society, professional bodies, and independent ombuds offices alongside regulators and technical assessors. Fifth, contextualisation and translation ensure that documentation, risk assessments, and notices are intelligible across languages, institutions, and levels of administrative capacity. The thesis maps areas of convergence and divergence between ubuntu ethics and a risk-based regulatory model, showing what can be retained without loss of legitimacy and where corrective adaptation is required. It specifies operational elements that move beyond rhetoric: calibrated sanctions combined with restorative remedies; capacity building through regional bodies; and documentation artefacts that support both transparency and practical oversight. Two conceptual applications test feasibility under contrasting conditions. In resource-constrained settings, participatory oversight and restorative duties strengthen weak enforcement. In more mature systems, relational standards deepen accountability and widen the evidence base for decision-making. The study concludes that an enforceable, risk-based approach, re-anchored in relational ethics, offers a credible path for African AI governance. It resists dependency, honours local moral reasoning, and contributes constructively to global debates on trustworthy AI. The contribution is both theoretical and practical: a clear set of principles, operational tools that regulators and institutions can adopt, and a route to interoperability that does not trade away contextual legitimacy. Masters 2026-04-23T12:19:33Z 2026-04-23T12:19:33Z 2026-03 Thesis https://scholar.sun.ac.za/handle/10019.1/136148 en Stellenbosch University 118 pages application/pdf Stellenbosch : Stellenbosch University |
| spellingShingle | Kalaba, Chiyesu Joseph AI Ethics in Africa: Postcolonial Challenges and Ethical Alternatives |
| title | AI Ethics in Africa: Postcolonial Challenges and Ethical Alternatives |
| title_full | AI Ethics in Africa: Postcolonial Challenges and Ethical Alternatives |
| title_fullStr | AI Ethics in Africa: Postcolonial Challenges and Ethical Alternatives |
| title_full_unstemmed | AI Ethics in Africa: Postcolonial Challenges and Ethical Alternatives |
| title_short | AI Ethics in Africa: Postcolonial Challenges and Ethical Alternatives |
| title_sort | ai ethics in africa postcolonial challenges and ethical alternatives |
| url | https://scholar.sun.ac.za/handle/10019.1/136148 |
| work_keys_str_mv | AT kalabachiyesujoseph aiethicsinafricapostcolonialchallengesandethicalalternatives |