Title: Governance of Artificial Intelligence in Insurance Operations: Extending IT Risk Registers and Board Reporting to Cover AI-Driven Decision Systems
Authors: Samuel Fadero
Stanley Nwakamma
Demilade Jooda
Serif Oyindamola Oyesiji
Volume: 10
Issue: 8
Pages: 1-16
Publication Date: 2026/08/28
Abstract:
This review examines the governance implications of artificial intelligence across contemporary insurance operations, with particular emphasis on extending established IT risk registers and board-reporting structures to address AI-driven decision systems. Using a structured narrative review approach, the study synthesises scholarly and institutional literature on AI applications, model risk, data governance, accountability, regulatory compliance, operational resilience, consumer protection, and executive oversight. The findings show that conventional technology-risk frameworks are often too static and asset-focused to capture the dynamic, probabilistic, data-dependent, and increasingly autonomous characteristics of AI. Material exposures include model drift, algorithmic bias, explainability limitations, data-quality failures, cybersecurity vulnerabilities, third-party dependencies, weak human oversight, and uncertainty surrounding responsibility for automated outcomes. The review further demonstrates that effective governance requires explicit model ownership, lifecycle validation, AI-specific key risk indicators, risk-appetite thresholds, continuous monitoring, independent assurance, and clear allocation of responsibilities across business units, risk and compliance functions, internal audit, executive management, and boards. Responsible AI also demands stronger safeguards for fairness, privacy, transparency, contestability, and consumer welfare, particularly where automated underwriting, pricing, and claims decisions may produce significant financial consequences. It also highlights the importance of locally responsive governance where regulatory capacity, infrastructure, and consumer vulnerability vary. The study concludes that AI risk should be embedded within enterprise governance rather than managed through isolated innovation structures. It recommends comprehensive AI inventories, strengthened board literacy, enhanced third-party controls, continuous assurance, human intervention for high-impact decisions, and proportionate sector-specific regulation. Future research should prioritise governance maturity, autonomous-agent accountability, comparative supervisory effectiveness, model interdependence, and responsible AI implementation across African and other emerging insurance markets.