Fraud involving mobile-money (MoMo) platforms in sub-Saharan Africa has increased significantly, notably as the use of mobile wallets has surged. Today, more than 60 percent of adults use mobile wallets to transact. What was once considered a convenient payment method has become a target for fraudsters. In the past two years, we’ve seen three dominant trends.
- Scammers use SIM-swap and SIM-jacking schemes to trick victims into losing their phones.
- Phishing via USSD-spoofed prompts that fool the user into providing their PIN as though “network maintenance” was taking place.
- Scams involve fraudsters posing as customer-service agents to harvest credentials.
AI-powered defenses are emerging, promising to deliver strategic breakthroughs. They include
training machine-learning models on millions of legitimate transactions. Some fintech platforms can now pinpoint anomalous behavior—whether it’s a flurry of high-value transfers at 2 a.m. or sudden jumps in transaction volume—with precision. In recent pilots, these systems have demonstrated over 90 percent accuracy in flagging truly suspicious activity, dramatically reducing both fraud losses and false alarms.
However, adoption hurdles remain. Many rural markets lack sufficient training data, which negatively impacts the quality of models. The convergence of biometrics and transaction histories raises privacy concerns. High computational costs also pose a challenge for smaller fintechs.
Painful experiences are increasing, especially among low-income communities.
- In Lagos, small trade is primarily reliant on PoS and mobile-money payments. The Police Special Fraud Unit recorded over 78,000 e-payment fraud cases in the past year. Individuals are reported to have lost an average of ₦500,000, equivalent to about a week’s sales for an average street-food vendor.
- According to a recent industry report, MoMo fraud cost Ghana over GH₵30 million in 2023. Market women and artisans suffered the most losses.
These shocks hurt supply chains and consumer trust. Many microenterprises revert to cash-only operations, undermining years of financial inclusion efforts.
What to do?!
- Regulators should encourage the sharing of anonymized public and private data for better AI models.
- Federated-learning frameworks are designed to ensure model training on mobile devices, thereby protecting user privacy while pooling insights from multiple devices.
- Quickly conduct secure experiments with unambiguous ethical boundaries through regulatory sandboxes for artificial intelligence (AI).
MoMo fraud, in short, is no longer a low-tech wedge against mobile inclusion. Instead, it’s become a high-end AI battleground. Africa’s best bet to stay one step ahead of fraudsters, who are ratcheting up the stakes, could be a timely marriage of strong machine-learning defences and smart regulation, or better cards.
