South Africans’ widespread caution born of high crime and low per‑capita income is helping to blunt the current impact of AI‑enabled fraud, but that edge may erode as more people go online, an industry founder has warned.
Limited incidence so far, but rapid change a risk
Alistair Pugin, founder of cloud technology enabler @Celyntis, told reporters that while South Africans are not immune to scams powered by artificial intelligence, local behaviour shaped by frequent exposure to crime makes many people harder targets for fraudsters. “Our level of gullibility is not as high as the rest of the world,” he said.
“We worry about everything; we look over our shoulders when we’re driving down the road. Our level of crime is a blessing in disguise.”
Pugin was commenting on figures from the Artificial Intelligence Incident Database (AIID). The database recorded 1 618 incidents between March 2023 and April 2026; ITWeb’s AI‑assisted analysis shows South Africa accounts for 12 of those entries, though one incident appears twice in the dataset.
| Metric | Value |
|---|---|
| Total AIID incidents (Mar 2023–Apr 2026) | 1 618 |
| South Africa incidents (AIID) | 12 (one duplicate) |
Those incidents reflect a wide range of harms, including fraud, manipulation, impersonation and the generation of false information. Pugin cautioned that the AI landscape is moving fast and that current protections tied to limited access and higher scepticism may not hold.
Why South Africa has been relatively spared
Pugin offered two linked reasons for the relatively low share of AIID cases from South Africa: a culture hardened by crime that fosters suspicion, and a lower gross domestic product per person that limits exposure compared with wealthier markets.
- Heightened vigilance: Regular exposure to crime can make people more cautious about unsolicited approaches online and offline, Pugin said.
- Limited access and affordability: Lower per‑capita GDP and still‑rising broadband and data affordability have, to date, constrained the number of targets available to operators of AI‑enabled scams.
However, Pugin warned those factors could shift. He pointed to the ongoing expansion of broadband, cheaper data and GDP growth as drivers that will put more South Africans online and potentially increase vulnerability.
He also placed the pace of AI change in broader context, referencing comments by Microsoft CEO Satya Nadella about computing capabilities accelerating faster than previous norms. Pugin said that ability now appears to be doubling roughly every six months — a speed he argued is unprecedented across industries — and that attackers will be able to exploit improvements in AI capabilities.
What this means for ordinary users
The immediate takeaway is not complacency. Even with relatively few recorded incidents, AI‑assisted scams can scale quickly once they become profitable. As connectivity grows, ordinary users who have relied on street smarts will increasingly face scams that leverage realistic voice synthesis, automated social engineering or personalised deepfakes.
Practical implications raised by the reporting include:
- the need for continued user education about AI‑driven impersonation and misinformation;
- the importance of affordable, credible reporting channels so victims can flag new scam types;
- the potential role of industry and regulators in tightening verification standards for digital communication.
While the AI Incident Database does not claim to be a comprehensive tally of every AI‑related wrongdoing, the numbers act as an early indicator of trends. For South Africa, that early signal is one of relative resilience today, with a clear risk that changing economic and connectivity conditions will alter the picture.
Keeping the balance will require not only technical counter‑measures from banks, telecoms and platforms but ongoing, plain‑spoken public education so South Africans can translate street smarts into digital caution.
