Strengthening AI Governance in Lending for Fair Credit Access in Kenya
Kenya stands at a pivotal moment in its journey to harness artificial intelligence (AI) for financial inclusion. However, without robust governance, the implementation of AI-driven credit scoring risks exacerbating existing inequities within the financial sector.
Current Landscape of AI in Lending
A recent Survey on Artificial Intelligence in the Banking Sector, conducted by the Central Bank of Kenya (CBK), revealed that half of Kenyan lenders have yet to adopt AI technologies. Among the 50% that do utilize AI, approximately 65% employ it for credit risk scoring.
Potential Benefits of AI in Financial Inclusion
AI has the potential to revolutionize risk assessment by utilizing digital footprints, transaction histories, mobile-money data, and behavioral analytics. This technological advancement could enhance access to credit for unbanked and underserved populations, particularly those operating within the informal economy.
The Need for Ethical Governance
Despite the promising benefits, experts warn that these advantages could turn detrimental without the establishment of ethical guardrails. The CBK survey indicates that few institutions employing AI have mechanisms for bias detection, algorithm explainability, or customer redress.
Denying credit based solely on an opaque algorithm is deemed not only unethical but also unwise. There is a pressing need for institutions to provide clear explanations for their credit decisions, especially in a society grappling with inequality and a widening digital divide.
Recommendations for AI Governance
Aligning Kenya’s National AI Strategy with CBK guidelines is essential. Financial regulations should explicitly incorporate principles of inclusivity, ethics, accountability, data integrity, and proportionate human oversight.
Regulatory sandboxes can be utilized to test AI systems for fairness and transparency, shifting the focus from mere predictive accuracy to equitable outcomes. Additionally, compliance functions must evolve; they should transition from post-development reviews to actively shaping AI systems from their design phases. Compliance teams are tasked with scrutinizing data sets, underlying assumptions, and optimized outcomes to mitigate risks and uphold ethical standards.
The Path Forward
The success of AI in credit risk management hinges on governance rather than technology alone. Kenya must critically assess whether AI will expand financial access or merely reinforce exclusion disguised as innovation. Achieving real progress requires regulators to establish enforceable standards while institutions must integrate meaningful AI governance into their operations. Without these measures, the promise of AI risks becoming a threat to inclusion rather than a catalyst for it.
