WHO AUDITS THE ALGORITHM?
AI GOVERNANCE: THE NEXT FRONTIER OF BOARD RESPONSIBILITY
Matthew R. Brown, Partner, EY
Artificial intelligence is no longer a future consideration for businesses. It is already influencing how organisations analyse data, make decisions, engage with customers and improve efficiency. Across South Africa and KwaZuluNatal, business leaders are exploring how AI can unlock growth, reduce costs and create competitive advantage.
Yet amid the excitement surrounding AI, an important question is emerging in boardrooms around the world: “Who is accountable when an algorithm makes the wrong decision?”
While much of the current conversation focuses on the opportunities presented by AI, far less attention is being given to the governance frameworks required to use these technologies responsibly. As adoption accelerates, AI governance is rapidly becoming one of the most important leadership and oversight challenges facing organisations today.
The Rise of AI in Business
Unlike previous technology advances, AI has the potential to influence decision-making itself. It can analyse vast volumes of data, identify patterns, generate content, assess risk and provide recommendations in seconds. Many organisations are already embedding AI into finance, supply chain management, customer service, risk management and operational processes. The benefits are significant. Improved productivity, enhanced insights and faster decision-making can help businesses navigate increasingly complex markets.
However, the power of AI introduces a new category of risk.
AI systems can produce inaccurate outputs, amplify biases in source data, generate misleading information or reach conclusions that humans do not fully understand. In highly regulated environments, these outcomes can have significant consequences for customers, investors, employees and broader stakeholders. As a result, organisations cannot view AI solely as a technology initiative. It must also be considered a governance priority.
Governance Must Keep Pace with Innovation
Historically, boards and executive teams have focused their oversight on areas such as financial reporting, cybersecurity, regulatory compliance and operational risk. Today, AI deserves a place alongside these priorities. The challenge is that many organisations have adopted AI tools faster than they have developed governance structures to manage them.
Questions that boards should be considering include:
How is AI being used across the organisation?
What decisions are being influenced or automated?
How is the accuracy and reliability of AI-generated information being assessed?
What controls exist to protect sensitive information?
Who is accountable for monitoring AI-related risks?
How do we ensure human oversight remains in place?
The answers to these questions will increasingly define how organisations build trust with stakeholders and regulators.
Trust Remains a Human Responsibility
One of the most important misconceptions about AI is that technology can replace accountability. While AI can support decision-making, responsibility remains firmly with people. Boards remain accountable for governance; executives remain accountable for business decisions and management remains accountable for the information used to support those decisions.
Customers, investors and regulators will not hold algorithms accountable. They will hold organisations accountable and this principle highlights why human judgement remains critical. Effective governance requires leaders to understand not only what AI can do, but also where its limitations exist. Professional scepticism, ethical decision-making, and informed oversight have never been more important.
Building a Responsible AI Framework
There is no single governance model that applies to every organisation. However, leading organisations are increasingly focusing on several common principles. First, transparency is critical. Businesses should understand where AI is being used and be able to explain significant AI driven outcomes.
Second, accountability must be clearly defined. Ownership for AI oversight should not sit solely within technology teams. Risk, compliance, legal, finance and leadership functions all have an important role to play.
Third, organisations should establish appropriate controls to monitor the quality, reliability and security of AI-generated outputs. Finally, AI governance should be viewed as an ongoing process rather than a once-off project. As technology evolves, governance frameworks must evolve alongside it.
AI-generated Outputs
Equally important is the validation of AI-generated outputs. While AI can significantly enhance efficiency and insight, its conclusions should never be accepted at face value. In both auditing and everyday business decision-making, professionals remain responsible for verifying, challenging, and corroborating AI-generated results. The effort required to perform this validation is often underestimated, creating a risk of overreliance on technology and a false sense of assurance.
As AI becomes more deeply embedded in organisational processes, maintaining professional scepticism and independent judgement will be critical to ensuring that the technology remains a trusted tool rather than an unquestioned authority.
The Opportunity Ahead
The organisations that will derive the greatest long-term value from AI are unlikely to be those that adopt it the fastest. Rather, they will be the organisations that combine innovation with trust, governance and accountability. This represents a significant opportunity for South African businesses.
By embracing responsible AI practices today, organisations can build confidence among customers, investors, regulators and employees while positioning themselves to capture the benefits of emerging technologies. Just as strong governance underpins confidence in financial reporting, effective AI governance will underpin confidence in the digital economy.
The conversation is no longer simply about what AI can do. It is about how organisations can use AI responsibly, transparently and ethically while maintaining the trust of the stakeholders they serve. In the years ahead, that trust may prove to be the most valuable competitive advantage of all.
