Words on Wealth: how AI is transforming financial services

Artificial intelligence (AI) is infiltrating our lives at a frightening speed. I can no longer watch an online video without questioning whether it is AI-generated or not. Is that really astrophysicist Neal deGrasse Tyson telling us the Earth is flat? No, it turns out that Tyson made and shared an AI-generated deepfake video of himself to highlight the potential dangers and misleading capabilities of the new technology. 

Thirty years ago, the internet revolution gave humanity an extremely useful communication tool – an ultra-convenient, ultra-wide-reaching media platform. It has been abused in countless ways: fanatics have used it to spread hate; organisations have used it to gather people’s personal information; fraudsters have used it to con people; pornography has proliferated. However, a rational person using the internet is still generally able to distinguish right from wrong and truth from falsehood. Regulators, always in catch-up mode, have had some success in tempering excess and providing a degree of consumer protection.

AI is different. In my view, its danger lies not in eventually subjugating humans, a favourite theme of science fiction; its danger lies in becoming indistinguishable from humans. And that, on the internet, is already happening. At some stage, the only way to tell whether a communication is genuine or not will be through direct human-to-human contact.

What is AI?

AI is an overarching term for computer systems that emulate human intelligence. The definition of AI by the Organisation for Economic Co-operation and Development has been widely adopted: “An AI system is a machine-based system that, for explicit or implicit objectives, infers from the input it receives how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments.”

At a fundamental level, AI is machine learning: the machine receives data from a wide variety of sources, identifies patterns in the data, and provides outputs such as deductions or predictions. It improves its performance based on its past performance – in other words, it learns through accumulating more data, including that gathered from its own experience.

Deep learning takes this further by imitating the neural networks in the human brain, gathering vast amounts of data and processing information on a number of levels, with minimal human intervention.

Generative AI, which we experience currently in the form of applications such as ChatGPT, generates new data that, among other things, can imitate the ways humans think, create and communicate – for example, the reproduction of human speech has advanced from a monotone “robotic” voice to models that accurately reproduce the natural inflections in the way we speak.

AI in financial services

AI is becoming increasingly important in banking, insurance and asset management. It takes much of the drudgework out of analysing facts and figures, boosts productivity and enhances the interface between consumers and providers. It is also increasingly being used to assess financial risk, monitor regulatory compliance and root out financial fraud. For example, AI is now being used by the South African Revenue Service to analyse large amounts of data quickly and accurately, enabling it to identify non-compliance.

The risks to consumers mainly concern the use of personal information and the generation of misleading information that can be detrimental to their financial wellbeing.

The Financial Sector Conduct Authority and the Prudential Authority recently published a report, “Artificial intelligence in the South African financial sector”, based on a survey of financial services providers regarding their adoption of AI and views on proposed regulation. The report reveals a steady increase in AI usage. Banks are at the forefront, with 52% of them actively employing AI, followed by payment providers (50%). 

Emphasising the need for ethical and responsible AI deployment, the report considers the need for robust governance frameworks, improved transparency, stronger consumer protection measures and increased consumer education around AI. Watch this space.

AI in investing

A few weeks ago, I wrote about the impact of AI on financial advice (“Financial advice, AI and the human touch” Personal Finance, September 20, 2025), suggesting that the human factor in providing advice should not be underestimated. But what about investing? Could AI eventually replace portfolio managers and beat them at their own game? Already we have sophisticated algorithmic trading that follows rules: if X is true, then buy or sell. 

It seems that active fund managers continue to see risks in letting AI make the investment decisions, although they are increasingly using it to refine their research. “The challenge when embracing AI is not a lack of data, but the frequent misinterpretation of it,” says Matis Mrazik, systematic investment specialist at London-based Jupiter Asset Management and portfolio manager of the Old Mutual Global Equity Fund, whose presentation I attended last week. 

“At Jupiter Asset Management, we are building on our experience in systematic investing by exploring how AI can enhance our research framework. This cannot be seen as delegating decision-making to black boxes or trusting back-tests run on historical data. In high-noise domains such as finance, collecting more data often translates to collecting more noise, unless one understands the structure and mechanisms that generated it,” Mrazik says. “We certainly see ourselves as practitioners of machine learning, with the caveat of approaching the field with discipline, as any scientist would. This applies to the data inputs we rely on and the subsequent modelling, acknowledging that market regimes change and structures evolve.”

PERSONAL FINANCE