Governing Agentic AI in Africa: A Systematic Literature Review of Ethical Challenges, Framework Gaps, and Ubuntu-Informed Governance Solutions
Abstract
Agentic AI systems pursue goals through multi-step planning, tool use, memory, and adaptive action, creating governance challenges that differ from those associated with conventional predictive or generative AI. This study systematically reviews governance evidence relevant to agentic AI in Africa and develops an Ubuntu-informed conceptual framework. Following PRISMA 2020, 91 records published from 2019 to April 2026 were analysed through quantitative content analysis and abductive thematic synthesis. The corpus comprises 12 agentic-AI-specific records, 3 agentic-adjacent records, and 76 broader AI-governance records used as contextual evidence. Five literature-derived gaps were identified: accountability across autonomous action chains, infrastructure-aware oversight, bias and representational harm, policy fragmentation, and community exclusion. These findings inform five proposed governance elements: communal impact assessment, collective consent and participation, infrastructure-contingent oversight, African dataset and bias auditing, and communal accountability. The framework is a literature-derived conceptual proposition, not an empirically validated governance instrument, and requires future case-study, expert-panel, and comparative policy testing.
Related
- AI-Driven Dynamic Trust in Zero Trust Architecture: A Systematic Literature Review of Machine Learning, Federated Learning and Governance Challenges
- The role of big data analytics in financial fraud detection: A systematic literature review
- Transparency, accountability, and scholarly integrity: A systematic narrative review of artificial intelligence governance frameworks in healthcare publication
- Hierarchical Cognitive Structuring and Game Mechanics Alignment in Educational Game Systems: A Systematic Literature Review
- Graph-Enhanced Knowledge Tracing for Explainable Learner Modeling: A Systematic Literature Review
BibTeX
@article{adeyelure2026governing,
title = {{Governing Agentic AI in Africa: A Systematic Literature Review of Ethical Challenges, Framework Gaps, and Ubuntu-Informed Governance Solutions}},
author = {Tope Samuel Adeyelure},
journal = {Journal of Information Systems and Informatics},
year = {2026},
month = oct,
doi = {10.63158/journalisi.v8i5.1738},
url = {https://doi.org/10.63158/journalisi.v8i5.1738},
note = {CC BY},
}
RIS
TY - JOUR
TI - Governing Agentic AI in Africa: A Systematic Literature Review of Ethical Challenges, Framework Gaps, and Ubuntu-Informed Governance Solutions
AU - Tope Samuel Adeyelure
PY - 2026
DA - 2026/10/24
JO - Journal of Information Systems and Informatics
DO - 10.63158/journalisi.v8i5.1738
UR - https://doi.org/10.63158/journalisi.v8i5.1738
AB - Agentic AI systems pursue goals through multi-step planning, tool use, memory, and adaptive action, creating governance challenges that differ from those associated with conventional predictive or generative AI. This study systematically reviews governance evidence relevant to agentic AI in Africa and develops an Ubuntu-informed conceptual framework. Following PRISMA 2020, 91 records published from 2019 to April 2026 were analysed through quantitative content analysis and abductive thematic synthesis. The corpus comprises 12 agentic-AI-specific records, 3 agentic-adjacent records, and 76 broader AI-governance records used as contextual evidence. Five literature-derived gaps were identified: accountability across autonomous action chains, infrastructure-aware oversight, bias and representational harm, policy fragmentation, and community exclusion. These findings inform five proposed governance elements: communal impact assessment, collective consent and participation, infrastructure-contingent oversight, African dataset and bias auditing, and communal accountability. The framework is a literature-derived conceptual proposition, not an empirically validated governance instrument, and requires future case-study, expert-panel, and comparative policy testing.
SN - 2656-5935
SN - 2656-4882
ER -