Leanne Mostert | Partner | Webber Wentzel | mail me |
Artificial Intelligence (AI) is no longer something financial institutions are preparing for. It is already embedded in how decisions are made, how products are designed and how risk is managed. Whether through fraud detection, algorithmic trading, customer engagement or compliance monitoring, AI is steadily reshaping the financial services sector.
Much of the conversation around AI still focuses on innovation and capability. It asks what these systems can do, how quickly organisations can adopt them and how they can improve efficiency. Beneath that, however, a more consequential conversation is emerging. It centres on ownership, accountability and governance.
AI and IP increasingly belong in the boardroom
Increasingly, that conversation belongs in the boardroom. AI does not operate in isolation. It sits at the intersection of technology, Intellectual Property (IP), risk and strategy. It creates value through data, models and automation. At the same time, it introduces difficult questions about ownership, protection and liability when things go wrong.
Many organisations are beginning to recognise that the real challenge with AI is not the technology itself. Instead, the challenge lies in whether existing governance structures can deal with the pace, complexity and legal uncertainty that accompany it.
The evolution of the AI debate reflects this shift. Much of the early focus centred on copyright risk. Generative AI systems train on vast volumes of data, much of which copyright protects. This has created increasing uncertainty around whether using copyrighted material in AI training may amount to infringement. Questions also remain about whether outputs generated by these systems may constitute infringement.
Copyright and the emerging AI and IP risk
In South Africa, the position is particularly nuanced. Copyright infringement does not depend on intention. Instead, it depends on whether someone has reproduced a substantial part of a protected work. This means that even ordinary workplace conduct can create legal exposure.
For example, employees may upload third-party research or sensitive client information into public AI tools. Such actions may expose organisations to legal risks. For many institutions, this has been a significant wake-up call. AI risk is not confined to developers or IT teams. It can arise quietly through everyday business practices.
Copyright, however, is only one aspect of the issue. The next question organisations inevitably confront is whether they can, in fact, own AI-generated innovation.
Patent law has started to provide some clarity. Courts in several jurisdictions have confirmed that AI itself cannot be recognised as an inventor. Human involvement remains central to patent protection. Ownership of an AI system does not automatically translate into ownership of everything it produces.
Protecting human ingenuity in AI
The organisations most likely to succeed in an AI-driven environment are not necessarily those using the most AI. Instead, they are those capable of clearly identifying, structuring and protecting the human ingenuity underpinning it.
AI-related inventions are not excluded from patent protection altogether. Significant opportunities remain for businesses that can demonstrate genuine technical innovation. This differs from merely layering AI onto existing processes. Patents, however, tell only part of the story. Some of the most valuable components of AI systems are precisely those that traditional IP frameworks struggle to protect.
Proprietary datasets, model refinement techniques, training methodologies and internal processes often occupy a grey area. They may prove too commercially sensitive to disclose publicly. Yet they may not always suit copyright or patent protection.
This is where trade secrets become critically important. For many financial institutions, confidential data and proprietary models rank among their most valuable assets. In many cases, they represent the true source of competitive advantage.
Trade secrets and the vulnerability of AI
Trade secrets are, however, inherently fragile. Their protection depends entirely on confidentiality. Once confidentiality disappears, so does the protection. AI can make that loss much easier.
Entering sensitive internal information into a generative AI platform may compromise confidentiality. This becomes particularly important where a platform retains or processes data outside the organisation’s control.
The same tools driving productivity and innovation may therefore expose the very information businesses are trying to protect. This tension explains why the conversation can no longer remain solely within legal, compliance or technology teams. AI and IP therefore need to form part of the same governance conversation. Organisations must understand not only what AI can produce, but also what intellectual property those systems use, create or potentially expose.
AI cuts across every part of an organisation. It affects how decisions are made, how risk is assessed, how products are developed and how accountability is allocated. It does not fit neatly within traditional corporate silos. Fragmented governance approaches are therefore unlikely to succeed.
Boards face a broader governance responsibility
Boards are being drawn into a far more active role. Directors do not need to become AI specialists, but AI increasingly shapes organisational risk, strategic direction and long-term value creation.
Effective oversight now requires more than understanding outcomes. Boards must also understand the assumptions, data and processes driving those outcomes.
AI governance is not ultimately about controlling technology. It is about ensuring that organisations remain in control of their decision-making, intellectual property and risk exposure as technology becomes increasingly embedded in business operations.
That requires different questions about AI and IP to be asked at leadership level:
- Is AI adoption aligned with business strategy?
- Does the organisation understand where its greatest AI-related risks lie?
- Are proprietary datasets, models and processes adequately protected?
- Is the organisation retaining ownership and control over the value AI is helping to create?
Governance must evolve with AI
AI is not static. Models evolve, outputs shift and risks change over time. Governance cannot therefore become a once-off policy exercise. Instead, it must remain continuous, adaptive and embedded into decision-making processes.
For financial institutions, this creates both pressure and opportunity. Organisations that approach governance as a compliance obligation may slow innovation while still struggling to manage risk effectively.
Those that approach it strategically are likely to move faster and with greater confidence. They can integrate intellectual property, governance and risk management into a coherent framework.
This is where AI and IP become particularly important. Bringing these considerations together can help organisations protect their assets while creating a more deliberate approach to AI adoption.
The conversation around AI is no longer solely about technology. It is about how organisations protect value, manage accountability and maintain trust in a rapidly changing environment. That is why this discussion has moved beyond innovation teams and legal departments.
It now belongs in the boardroom. AI and IP are therefore no longer separate considerations. They form part of a broader governance challenge that boards and financial institutions must address as AI becomes increasingly embedded in business operations.
