Nearly 69% of biometric fraud attempts targeting African fintech platforms now involve artificial intelligence. That figure, drawn from a study of more than 200 million identity verification checks across 35 African countries, is not a forecast — it is what happened in 2025. For South African businesses operating in fintech, banking, or any sector that verifies identity online, the implications are significant.

The Threat Has Moved Past the Front Door

For years, fraud prevention focused on account creation. The logic was straightforward: if you could verify someone's identity when they signed up, you were protected. Most platforms deployed document checks and selfie verification at onboarding — and that was largely the end of it.

Fraudsters have adapted. According to the Smile ID Digital Identity Fraud in Africa 2026 report, account takeover attempts — targeting login, password reset, and device-change events — are now five times more common than fraud at account creation. The security controls at these later touchpoints are often less rigorous, and criminals know it.

The Scale Is Staggering

The report documents one case that illustrates the industrial scale of modern identity fraud. A single fraud network used approximately 100 stolen biometric identities to generate more than 160,000 verification attempts across several African fintech platforms in a single month. Some of those stolen identities were used in over 12,000 separate access attempts.

This is not opportunistic fraud. It is organised, systematic, and increasingly automated. The report describes the emergence of supply chains for digital identities — criminal ecosystems that harvest stolen credentials from hacked databases, package them, and deploy them at scale using AI-powered tools.

AI Has Made Fraud Cheaper and More Convincing

The 69% figure refers specifically to AI-enabled biometric fraud: deepfakes, synthetic faces, and automated facial manipulation techniques that can fool standard liveness detection systems. These tools are not expensive or technically complex to obtain. They are widely available on underground markets and continue to improve.

West Africa saw retail banking fraud rise approximately 50% in 2025, driven largely by AI-assisted account takeover attempts. The pressure is spreading south. Any platform that relies on biometric verification without continuous, AI-aware liveness and behavioural checks is operating with a gap that criminals are actively exploiting.

Why This Threatens More Than Just Money

Financial inclusion across Africa has been one of the continent's major achievements of the past decade. The share of adults with a financial account has risen from 34% to nearly 60%, representing over 200 million people brought into the formal financial system. That progress depends on trust.

As Smile ID's CEO Mark Straub noted in the report, if users begin to perceive cash or informal channels as safer than digital finance, the gains of the past decade risk being reversed. Fraud is not just a cost of doing business — it is a threat to the infrastructure of digital economies.

What Businesses Need to Do Now

Addressing this threat requires more than upgrading your onboarding flow. It requires rethinking identity verification as a continuous process, not a one-time event. Specifically:

• Verify at every critical touchpoint — not just sign-up. Login, password reset, and device changes are now primary attack surfaces.
• Use liveness detection that is AI-aware. Basic selfie matching is no longer sufficient against deepfake-enabled fraud.
• Monitor for behavioural anomalies. When a single identity generates hundreds of verification attempts, something is wrong — your system should detect it automatically.
• Integrate document verification with biometric matching. Cross-referencing identity documents with real-time government databases reduces the value of stolen credentials.
• Treat verification as infrastructure. Fraud prevention should not be a bolt-on feature — it should be embedded into every identity-sensitive interaction your platform handles.

The Bottom Line

The data is clear: AI-powered identity fraud is no longer an emerging risk in South Africa and across Africa — it is the current reality. Businesses that continue treating identity verification as a box to tick at onboarding are exposed. Those that build continuous, AI-resilient verification into their operations are not.

WhoYou manages the complete digital identity lifecycle, from onboarding verification through to ongoing authentication — giving businesses the tools to stay ahead of a threat that is evolving faster than most security teams can track.