The language of ethical AI has become reassuringly familiar. Organisations promise fairness, transparency, inclusion and human oversight. Each principle matters, yet, without a standard against which conduct can be tested, these principles risk becoming values that everyone endorses but no one is held to.
In financial services, the test must be more demanding because the decisions we make reach directly into people's lives. A model may influence whether someone receives credit, how their insurance risk is priced, or what advice reaches their household. For the institution, this may appear as a prediction. For the person affected, it can mean an opportunity opened—or a door firmly closed.
Our position is straightforward: wherever AI shapes a decision that matters to a person's life, whether that person is a customer or an employee, we should ensure they can understand how the outcome was reached, have access to meaningful recourse where concerns arise, and trust that accountability sits firmly with the institution, not just the algorithm.
Within Old Mutual, clear principles and practices guide how AI is designed, deployed and overseen, so that every system we use remains trustworthy, accountable, and aligned with our values, our risk appetite and our strategic objectives. Governance of this kind is not a brake on innovation; it is what allows innovation to be trusted.
South Africa's financial authorities are already asking the right questions. In May, South African Reserve Bank Deputy Governor and CEO of the Prudential Authority, Fundi Tshazibana, signalled that financial institutions remain accountable for the explainability of decisions, even when those decisions are informed by AI.[1] We agree, and would go further; an explanation has limited value if the person affected cannot do anything with it.
Meaningful explanation does not mean exposing source code or compromising security. It means giving a customer an answer they can understand and act on: what information shaped the decision, and what could lead to a different outcome. An institution may choose to buy the model, but it cannot outsource accountability for the decisions made with it.
The same test belongs inside our own walls
As Group Chief Human Capital and Corporate Affairs Officer, I cannot confine this argument only in relation to customers. Similar, AI-enabled systems are increasingly being used in Human Capital processes globally. Increasingly, the most interesting AI use cases in Human Capital are no longer in strategy or reporting. They are in the operational engine room of HR: recruitment, onboarding, service delivery, payroll, learning, workforce planning, employee relations, and manager support.
When technology influences decisions that affect a person's career, livelihood, or future opportunities, the standards of fairness should be no lower than those we demand in serving our customers. Every employee deserves a decision that can be explained in clear and understandable terms, a meaningful avenue to question or challenge the outcome, and the confidence that accountability ultimately rests with a human being. Algorithms may inform decisions, but they should never replace human judgment, oversight, or responsibility. In an AI-enabled workplace, transparency, fairness, and human accountability are not simply ethical considerations; they are essential to preserving trust.
This cannot be treated as a peripheral concern or delegated to a policy annexure. The ethical standards we apply externally must also govern how we treat our own people. Internal and external ethics are not separate disciplines; they are the same principles applied to different relationships. An institution that protects its customers while neglecting its employees will quickly find its credibility weakened on both fronts. A workforce that does not trust how decisions are made, how performance is assessed, or how opportunities are allocated will struggle to build the trust, confidence, and advocacy upon which customer relationships ultimately depend. Trust is indivisible. If fairness, transparency, and accountability are absent inside the organisation, they will inevitably be questioned outside it as well.
Ultimately, the responsible use of AI is not a matter of compliance alone. It is a reflection of organisational character. The organisations that succeed in the AI era will be those that use technology to strengthen human dignity rather than diminish it, combining the power of intelligent systems with the judgment, empathy, and accountability that only people can provide.
Africa must not automate its blind spots
Africa has an opportunity to engage with AI in a way that is both pragmatic and purposeful, using it not as an end in itself, but as a means to expand access, improve productivity, strengthen service delivery, and unlock human potential. While it would be premature to make definitive predictions about the pace or scale of adoption, there is little doubt that AI will increasingly shape how organisations serve customers, develop talent, make decisions, and compete in a rapidly evolving global economy. The question is not whether AI will influence our future, but how intentionally and responsibly we choose to apply it.
That responsibility begins with recognising that trust is the foundation upon which sustainable adoption rests. The benefits of AI will only be realised if employees and customers have confidence that the systems affecting them are fair, transparent, secure, and accountable. This requires clear governance over data use, meaningful human oversight of consequential decisions, protection against bias and unintended discrimination, and robust safeguards for privacy and security. Technology may augment human capability, but it should never diminish human dignity.
For organisations across Africa, the challenge is therefore not simply to adopt AI, but to adopt it wisely. We should welcome the opportunities it presents to enhance customer experience, broaden financial inclusion, improve workforce effectiveness, and remove administrative friction. At the same time, we must remain disciplined in protecting the rights and interests of the people we serve and employ. The most successful institutions will be those that strike the right balance: embracing innovation with ambition while governing it with care.
Celiwe Ross is Group Chief Human Capital and Corporate Affairs Officer at Old Mutual. She will be speaking at the Singularity Summit 2026 at Sandton Convention Centre on Wednesday 21 and Thursday 22 October. Presented in collaboration with Old Mutual, event sponsors Google Cloud and Industrial Development Corporation (IDC) and metaverse partner UBU, this year’s Summit will bring together over 40 speakers from around the world to examine the technologies, ideas and shifts influencing the future of business and society. For more information, or to book, go to https://singularitysouthafricasummit.org/