FINANCE FEATURE: The AI code war-Africa’s digital economy is under threat

by | Sep 29, 2025

  • The rapid development of artificial intelligence (AI) correlates with the evolution of digital crimes.
  • Financial institutions use AI to detect fraud, but criminals are also using AI to develop new ways of committing fraud.
  • In the pursuit of developing world-class AI for Africa, data security is the most critical element in securing Africa’s digital economy.

In an undeniable reflection of Africa’s burgeoning digital economy by Mastercard, the continent’s AI market is projected to skyrocket from an estimated R78 billion in 2025 to a massive R286,9 billion by 2030. South Africa’s projected AI market size is over R20 billion as of 2025, according to the report by Mastercard. As a direct response to this financial surge, nations like South Africa are moving to secure this digital future.

As we progress further in the development of a digital Africa, traditional, rule-based security systems are proving to be slow and ineffective. From phishing emails, spam texts to sophisticated deepfake attacks, criminals are constantly evolving their methods.

Standard Bank’s Head of payments, Rufaida Hamilton, wrote an article discussing the role of AI in detecting, monitoring, and preventing payments fraud. Hamilton states that AI is transforming fraud prevention in the financial sector by moving beyond traditional rule-based systems to proactively detect and monitor payments fraud in real time. AI-powered systems can detect these patterns in transaction data, identify unusual customer behaviour, and even analyse text and voice to flag suspicious activity in real-time.

Above: Data chart showing the number of AI-related scams over a three-year period .
Graphic: Katlego Makhutle/TRM Labs

However, the article also notes that this is a continuous cat and mouse game, as criminals are simultaneously using AI to develop more sophisticated scams and mimic human behaviour. The South African Banking Risk Information Centre (SABRIC) released an annual report in 2024 stating that there’s an uptick in AI-powered financial crimes, with criminals getting more creative in their use of AI to commit phishing, deep fakes and creating synthetic identity frauds.

According to SABRIC, a major underlying threat is the rise of synthetic identities, which are fictitious personas created by AI that blend real and fake data. These digital phantoms are used to apply for loans and open fraudulent accounts, bypassing traditional verification systems designed to spot real people.

Application fraud remains a significant concern, with false applications alone contributing 2.9% of card fraud losses. Beyond just tricking individuals, AI is also enabling a new kind of “ghost” fraud that attacks the very foundations of the financial system. TRM Labs also reported that criminals are now using AI to generate hyper-realistic deepfakes of executives to trick employees into wiring millions.

SABRIC’s report also reveals that criminals are no longer just relying on clumsy, rule-based attacks but are harnessing AI to craft a new generation of scams. Gone are the days of misspelt phishing emails and awkward grammar.

Instead, criminals are deploying generative AI to produce “error-free phishing emails” and “AI-generated WhatsApp messages.” This shift makes every text and email a potential trap, turning the most common forms of communication into potential tools for digital fraud.

While overall financial crime losses saw a decline of nearly 18%, dropping from R3.3 billion in 2023 to R2.7 billion in 2024, digital banking fraud has surged, becoming the dominant threat South Africa’s digital ecosystem.

The number of reported digital fraud cases more than doubled, soaring from 31,612 in 2023 to 64,000 in 2024. This dramatic increase resulted in a proportional rise in financial losses, which climbed from R1 billion to over R1.4 billion in the same period.

The fight for control over Africa’s digital financial landscape is a new kind of “Code War,” where fintech innovations and AI-powered defences are locked in a continuous escalation with sophisticated digital criminals.

However, the adoption of AI in this context is not without its own set of challenges, particularly concerning data privacy, algorithmic bias, and the need for robust regulatory frameworks. The solution to these escalating threats extends beyond technology and connects directly to the concept of data sovereignty and security in Africa.

Wits University’s School of Electrical and Information Engineering lecturer, Dr. Martin Bekker, noted that AI in Africa still needs to be given representative data by Africa and for Africa. Bekker highlights that sharing data with any Language Learning Model (LLM) is not secure or private and sharing data with these tools does not present any “unique” security challenges; instead, it exacerbates existing challenges.

“Digital security is always relative [and] never absolute…[its] a bit of a cat-and-mouse game. However, there are ideas such as hosting your own open-weight LLM internally, as opposed to using an online service, which are emerging as best practices. As for the ethics – AI training appears to rely on quite a bit of IP misappropriation, super-high energy use, and if there is RLHF (Reinforcement learning from human feedback), possible labour exploitation too,” Bekker stated.

At the recent GovTech 2025 conference, South African Minister of Science, Technology and Innovation, Professor Blade Nzimande, warned that without digital sovereignty, the country’s national sovereignty is at risk, as data becomes increasingly controlled by foreign tech giants.

“We need digital sovereignty. We can’t have our data controlled by everybody, anywhere in the world,” Nzimande stated. The purpose of building local data centres is not solely based on the physical storage of data. It also points to a strategic move that is working to ensure that African-based financial data remains within the continent’s borders, is subject to local laws and is used to develop a domestic AI industry that serves African needs and values.

The long-term defence against AI-powered crime and data breaches must be a holistic one. However, simply localising data and regulating financial institutions does not automatically guarantee security or sustainability. The story of AI in Africa’s financial sector is therefore a complex narrative of innovation and vulnerability. An ever-changing and unfolding story, where advanced digital technology is the best defence against crime and the most potent weapon for digital fraudsters.

FEATURED IMAGE: Hand holding phone with scam alert on the screen Photo: Katlego Makhutle

RELATED ARTICLES: