On April 29 2026, Strathmore University, in collaboration with ABSA Bank Kenya,
hosted a conference on “Ethical AI in Finance: Driving Inclusion, Trust, and
Innovation in Kenya”. DIFA’s Principal Consultant, Ikran Ali Abdirahman, participated
in the keynote panel examining the impact of data governance, AI ethics, and
regulation on Kenya’s financial sector. She discussed responsible AI in digital credit
and the importance of promoting fair, inclusive, and transparent lending practices in
Kenya.
The session highlighted that as AI continues to transform Kenya’s financial sector,
digital lenders are increasingly using alternative data such as transaction behaviour,
airtime top-ups, app usage, device metadata, and location patterns to expand
access to credit. While these technologies promise greater financial inclusion, they
also present serious risks if not designed and governed responsibly.
A case study from the DIFA Responsible AI Toolkit for African Startups, highlights
this concern through a fictional case study of fintech company using AI-driven credit
scoring models, showing the “digital visibility bias”, that digital activity does not
always reflect true creditworthiness, among other risks. The key consideration
arising from this case study was the importance of embedding responsible AI
governance from the design stage throughout the whole AI lifecycle.
Financial institutions should ensure AI systems do not reinforce socio-economic
inequalities through indirect discrimination linked to factors such as location, device
type, or transaction frequency. Measures such as representation checks before
model training, limiting unnecessary data collection, and continuous monitoring for
bias and performance disparities are essential. Privacy and data protection also
remain central concerns. Since AI systems process sensitive personal information,
organisations should adopt “minimum necessary data” principles, encryption,
pseudonymisation, restricted data access, and clear retention policies. Institutions
should also conduct Data Protection Impact Assessments (DPIAs), maintain
transparent customer communication, and ensure accountability for AI outcomes
remains with the deploying institution.
As highlighted during the session, the DIFA Responsible AI Toolkit ultimately calls
for proactive and inclusive AI governance within Kenya’s financial sector. Rather
than waiting for stricter regulation, financial institutions are encouraged to proactively
invest in AI literacy, customer-centred AI policies, and robust risk assessment
frameworks to uphold their fiduciary duties.
Kenya’s success with Safaricom’s mobile money solution M-Pesa, demonstrates that
technology can drive meaningful inclusion when deployed responsibly. The future of
AI in digital lending therefore depends not only on innovation, but on ensuring
systems remain fair, transparent, inclusive, and respectful of data protection rights.
Swahili 
English