Tawanda Musarurwa
A FEW examples from around the world have shown that although artificial intelligence (AI) can offer solutions to some challenges in the insurance industry, the technology comes with significant risks.
To illustrate the positive aspects, as far back as 2018, Indian insurer ICICI Lombard General Insurance adopted an AI-based cashless claims settlement system to verify health insurance claims, as the company sought to authorise cashless claims within 60 to 90 seconds and reduce turnaround time for investigations to within 24 hours.
On the other side of the coin, in 2019, Lemonade, an AI-driven insurance company, faced criticism when users accused its algorithm of potentially discriminatory practices in approving claims and setting premiums.
Broadly, a number of observers say AI can help streamline operations, improve risk management and enhance customer experiences within the increasingly complex insurance sector.
Mugonat Systems CEO Mr Elvis Mugovi believes AI-driven systems can reduce instances of insurance fraud.
“Insurance fraud significantly impacts Zimbabwe’s insurance sector, with fraudulent claims estimated to account for around 30 percent of payouts annually,” he said.
“By automating processes like claims handling and underwriting, AI enables rapid analysis of extensive datasets, uncovering anomalies and suspicious patterns often missed by human agents.”
The local software firm is currently working to incorporate AI technologies into its existing solutions for medical aid and short-term insurance providers.
Effective regulation has the potential to steer the local insurance sector towards innovation, while safeguarding public trust.
The Insurance and Pensions Commission (IPEC), which regulates the local insurance sector, will need establish a sector-specific AI policy and tighten regulatory oversight over insurers’ adoption of AI.
IPEC has said it is developing a cyber security framework, although it is not yet clear if broader AI issues will be factored into that policy framework or a separate AI policy will be developed.
“To curb incidences of policyholder data and information abuse by insurers, the Commission is developing a cyber security framework,” said IPEC director insurance and micro-insurance Mrs Sibongile Siwela.
“The regulatory framework will set out specific requirements for insurers to implement robust security measures to protect personal data from unauthorised access or disclosure.
“Insurers are expected to adopt appropriate technical and organisational measures to safeguard data integrity, confidentiality and availability.”
Although Zimbabwe still lacks concrete regulatory frameworks on the use of AI, a robust AI policy for insurers is critical for addressing the sector’s inefficiencies and leveraging opportunities.
Parameters are not the enemy of progress – they may actually turn out to be the compass, as many insurers appear to be stuck in the “pilot phase” of implementing AI systems.
The country’s insurance sector faces unique challenges, including low penetration rates, operational inefficiencies and widespread mistrust.
According to statistics from IPEC, the country’s insurance penetration rate is around 1,6 percent this year, down from 1,8 percent in 2023, which points to significant untapped markets.
Comparably, the global average insurance penetration rate is around 7 percent.
The local insurance sector also saw a gross premium written (GPW) growth of -10 percent in 2023.
Operational inefficiencies such as manual processes and fraudulent claims continue to drain resources, while many Zimbabweans distrust insurers, citing opaque practices and poor customer service.
A functional AI policy for the sector, backed by robust regulations, could be a game-changer.
AI can directly address a number of challenges in the sector.
For instance, chat-bots could provide policyholder assistance in the country’s numerous local languages around the clock.
Given that Zimbabwe has 16 official languages – Chewa, Chibarwe, English, Kalanga, Koisan, Nambya, Ndau, Ndebele, Shangani, Shona, Sign Language, Sotho, Tonga, Tswana, Venda and Xhosa – it is fair to say that most insurers’ customer service is currently not inclusive.
Predictive models could also design affordable micro-insurance products for rural populations.
And fraud detection algorithms could flag suspicious claims, saving millions.
However, unlocking AI’s potential requires policies that foster innovation, while protecting consumers.
Experts say there is need for, at least, a national AI regulatory framework.
Discussing during a podcast themed, “Reimagining insurance with a comprehensive approach to gen AI” in August, global management consultants McKinsey and Company partner Mr Khaled Rifai said:
“In terms of regulation in Europe, the EU Artificial Intelligence Act has recently been passed.
“With room for national regulations, national regulators of the insurance industry will look at certain cases to determine standards. In my experience, the regulations are good enough for clients to work with.
“I wouldn’t start with high-risk cases concerning decisions that impact the life and health of the insured, but instead begin with other use cases that we’re certain we can implement in a secure, customer-friendly way.
“The thing to remember is that nothing is static, and the ongoing process of shaping regulations means taking things one step at a time.”
Government’s national AI policy framework, announced last month (but yet to be promulgated), is a step forward.
And because data protection is paramount in the use of AI systems, Zimbabwe’s Data Protection Act (2021) is a good foundation, but it must continually evolve to address AI-specific challenges.
In a significant development, in September Government promulgated Statutory Instrument 155 of 2024, which set a regulatory framework for licensing data protection entities and appointing data protection officers in the country.
Muvingi and Mugadza Legal Practitioners say firms that handle personal data must now align their data protection practices with the new regulations.
“The introduction of SI-155 of 2024 will have a significant impact on organisations handling personal data,” says the law firm.
“They must align their data protection practices with the new regulations, appoint qualified data protection officers and obtain the necessary licenses to operate legally.”
At a broader level, some observers say there is need to align the national AI strategy with associated strategies at the regional and continental levels.
Development specialist and Media Institute of Southern Africa (MISA) regional director Dr Tabani Moyo has called for the prioritisation of bottom-up processes to shape the African Union (AU) Artificial Intelligence blueprint titled: Continental Artificial Intelligence Strategy: Harnessing AI for Africa’s Development and Prosperity’s uptake and diffusion across the continent.
Dr Moyo was speaking at the Organisation for Economic Co-operation and Development-Africa Union meeting this month (November 2024) in Cairo, Egypt.
“We researched the State of the AI Policy Regulation Framework in Southern Africa year-long.
“Our observations and outcomes show a need for structured linkages between the African Union, its regional economic communities (RECs) and the member states if the AI continental blueprint is to gain traction and ownership at the member states level,” he said.
“This is mainly because the member states see the RECs as their immediate reality compared to the continental body, which makes the blueprint more abstract if there is no mechanism for structuring pathways and runways with the regional blogs to inform the context at national levels.”
What insurers need to do
Insurers must ensure AI systems are transparent, fair and ethical.
Addressing the Insurance Institute of Zimbabwe (IIZ) annual conference earlier this month, Mrs Siwela highlighted a number of existing unethical conduct by insurers, including unreasonable delay of payments; denying a policyholder’s claim despite overwhelming evidence to support it, and mishandling of clients’ information.
These misconducts can either be worsened or eliminated by AI systems.
Insurers need to ensure the latter.
Customers denied claims by algorithms should understand why and have the right to appeal.
The companies should ensure that algorithms should not perpetuate biases, such as charging higher premiums to rural residents or women, based on skewed historical data.
The macro perspective
From a macro and policy perspective, data access must be balanced with privacy, and offering tax incentives for AI investment can accelerate adoption.
An independent oversight body is essential to certify AI systems, monitor compliance and address grievances.
And educating consumers about AI’s benefits and safeguards will also be crucial to building trust.
Zimbabwe can draw lessons from other countries.
In China, insurers like Ping An use AI for fraud detection and customer service under strict government oversight.
In India, insurers use AI to assess crop damage remotely, providing faster claims for farmers.
Closer to home, South Africa’s Protection of Personal Information Act ensures fairness in AI-driven decisions, enabling innovations like Discovery Health’s dynamic premium adjustments.
Local insurers could adopt similar tools to extend services to underserved rural communities.
The potential rewards are significant.
AI could make insurance more accessible, boosting penetration rates while reducing operational costs.
Transparent and fair AI systems could rebuild consumer confidence in the country’s insurance industry, which has viewed with skepticism, especially after the value loss that occurred during the currency conversion in 2009.
For Zimbabwe’s insurers, AI is not just a tool for catching up – it can be a means to leapfrog into the future.
With the right policies, the country could emerge as a regional leader in AI-driven insurance, reaping both economic and social benefits.
ORIGINALLY PUBLISHED IN THE SUNDAY MAIL – https://www.heraldonline.co.zw/policy-regulation-missing-links-for-local-insurance-sectors-ai-adoption/