Before setting rules, insurance regulators must live with AI tools long enough to understand their real-world behaviour, says the head of the supranational body setting insurance rules across 14 mostly Francophone West and Central African countries.
Machine learning, AI and GenAI are poised to transform insurance distribution and risk modelling.
South Africa’s Old Mutual has highlighted AI’s potential to personalise health and life insurance premiums using fitness tracker data. Nigeria’s Leadway Assurance is deploying VehiScannerAI, allowing policyholders to scan damage to their vehicles to accelerate motor insurance claims, while Kenya’s Britam is backing AI insurance startups via its BetaLab accelerator.
‘New challenges to consumer protection’
Speaking at AFIS 2025, Blaise Abel Ezo’o Engolo, Secretary General of the Inter-African Conference on Insurance Markets (CIMA), said AI was developing “at great speed” and was “the burning subject” in the insurance sector.
“Artificial Intelligence, digitalisation and Insurtech create new challenges to consumer protection,” said the regulator.
He added that insurance customers with weak digital literacy could be exposed to opaque contracts embedded in digital apps, algorithmic bias, and data confidentiality issues that transcend national borders.
‘Regulating AI too late or too early may be harmful’
“Most of our rules have been set in a way that served only traditional elements,” said Ezo’o Engolo. “We have to adapt to new things.”
But the CIMA Secretary General cautioned against moving too quickly. “Regulating AI too late or too early may be harmful,” he said.
“Putting in place a regulation is one thing; supervising and controlling it is another. In a scheme where activity precedes regulation, it is necessary to understand and live with these new technologies before pretending to regulate or create rules,” he continued. He added that guidelines or parliamentary acts must be precise enough to keep pace with technological evolution and should be supported by skilled personnel and robust monitoring systems. “This calls upon us to revise the control systems,” he said.
Four of 55 African states have a national AI policy
AI regulation in Africa has been gradually evolving since the 2013 Smart Africa Manifesto.
Yet among 55 African states, just four have a national AI policy and seven have an AI strategy. Fifteen run AI working groups, while 41 have privacy and data protection laws.
“This shows that African countries do not yet have enough national or regional laws, but they are still doing their best to be active and proactive,” said the CIMA chief.
The African Union’s 2019 Sharm El Sheikh declaration called for a common African stance on AI and the creation of an AI working group. This led to the 2021 Blueprint – Artificial Intelligence for Africa, which offered recommendations for national strategies, and in July 2024 the AU unveiled its Continental Strategy on AI.
“Through parliament and other instances, we must take this into account. This should be translated into texts—decrees, circulars, regulations, directives, and so on—to regulate this domain,” said CIMA’s Secretary General.
The upcoming African Financial Industry Barometer by Deloitte and AFIS, due January 2025, found insurers see AI’s main applications in personalising products, streamlining processes, and simplifying onboarding.
