More than half of SA firms suspect AI-powered fraud attempts

Artificial intelligence is being used against South African companies, and many of them feel poorly equipped to respond. The Citizen reported on research showing that over half of local businesses have either come across AI-enabled fraud or suspect that they have. The tactics include deepfakes, cloned voices and convincing phishing scams.

Despite this, fewer than one in four businesses believe their identity verification systems are highly effective at stopping fraud.

What the findings show

According to the research, 52.2% of businesses have seen, or suspect, attempts that used AI-written phishing messages, imitated voices or deepfake technology. Only 44.2% say they are confident they could spot fraud assisted by AI.

AI-driven identity fraud now tops the list of concerns. About 76.6% of organisations call it a key risk that needs urgent attention. Two years ago that figure was 59.2%.

Confidence in existing defences has moved in the opposite direction. The share of businesses describing their identity verification as highly effective stands at 22.9%, down slightly from 24.5% before.

Synthetic identities are also on the radar. These are fake profiles built by mixing real details with invented ones. Roughly two in three businesses are preparing for this kind of threat, and worry about deepfake fraud has grown noticeably.

Gaps in preparation

Many firms would struggle if an attack succeeded. Fewer than a quarter have formal, tested procedures for dealing with biometric fraud or a compromised identity. More than one in three have no response plan at all.

Small and medium-sized companies reported less confidence in their fraud defences than large enterprises did. This suggests that money and technology make a real difference to how ready a company is.

There are signs of a reaction. Over half of the businesses surveyed intend to spend more on tools that verify identities and prevent fraud during the next two years, as they try to keep pace with increasingly sophisticated AI scams.

What this means for South Africans

The findings suggest that fake messages and imitated voices are becoming a growing part of the fraud landscape. Customers and staff may be the targets, as well as the companies themselves. Because businesses plan to invest more in verification, people may find that proving who they are becomes more thorough in the coming years. Smaller firms, which reported lower confidence, may be where gaps are most likely to remain.

This story is based on reporting by The Citizen.