Mobile money in 2026: fraud controls are now financial infrastructure

Mobile money has crossed another scale threshold

The GSMA State of the Industry Report on Mobile Money 2026 says the global ecosystem processed $2.1 trillion in transactions in 2025, with 2.3 billion registered accounts and 593 million active 30-day accounts. The report also notes strong growth in merchant payments, which reached $155 billion in 2025. For Sub-Saharan Africa, where mobile money is daily infrastructure, these figures are not only an inclusion success story. They describe a systemic risk surface.

When millions of users pay, receive wages, borrow and trade by phone, fraud is no longer only a customer-level loss. It can affect trust in operators, agent liquidity, small-business continuity and the perceived reliability of private and public digital services.

Red Card 2.0 shows fraud professionalisation

INTERPOL announced on 18 February 2026 that Operation Red Card 2.0 led to 651 arrests across 16 African countries and recovered more than $4.3 million. The operation, conducted from 8 December 2025 to 30 January 2026, targeted high-yield investment scams, mobile money fraud and fraudulent mobile loan applications. INTERPOL said investigations exposed scams linked to more than $45 million in losses and identified 1,247 victims.

The examples cited by INTERPOL match African market realities: fake mobile loans in Côte d'Ivoire, investment fraud in Kenya, phishing and identity theft in Nigeria, and compromised staff credentials at a major telecom provider. This is not only a story of isolated low-level scammers. It combines social engineering, digital infrastructure, online accounts, SIMs, apps and sometimes internal access.

AI improves attacker productivity

Generative AI does not create all fraud, but it improves its economics. It can produce more convincing messages in French, English and local languages, generate fake documents, personalise call scripts, create coherent social profiles and quickly test multiple versions of a lure. The risk is acute for quick loans, investment promises and customer-support impersonation.

Synthetic voice adds another layer. A customer may receive audio that appears to come from an agent, employer or family member. Fraud becomes harder to detect through old cues such as poor spelling or awkward phrasing. Controls therefore need to move toward transaction evidence: device reputation, account history, recent SIM changes, transaction velocity, geographic consistency, malicious-app intelligence and agent behaviour.

Why Africa should treat this as infrastructure

In many African markets, mobile money replaces or complements banks for essential activity. A large fraud wave can therefore have disproportionate social impact: loss of savings, distrust of digital channels, a shift back to cash, pressure on support centres and regulatory intervention. The response cannot simply be to tell users to be careful.

Operators, banks, fintechs and regulators need stronger signal sharing. A fraudulent domain, malicious APK, mule account, phone number and IP address have limited value if they remain inside one organisation. They become useful when shared quickly in an operational format, with data-protection rules and a process for correcting false positives.

Practical priorities

• Build risk scoring that combines account, device, SIM, agent, channel and transaction behaviour.
• Harden SIM change, PIN reset and agent onboarding procedures.
• Monitor Android apps and fake loan forms abusing operator brands.
• Create a shared indicator feed: domains, numbers, mule accounts, APKs, hashes and message templates.
• Train support teams for synthetic voice and executive impersonation scenarios.
• Publish short anti-fraud warnings tied to real campaigns observed in the market.

The regulator's role

Regulators should avoid two extremes: leaving each provider to fight fraud alone, or imposing rigid controls that make legitimate access painful for low-income users. The useful role is to define minimum controls, require measurable incident reporting, encourage shared intelligence and test how providers handle mass fraud campaigns. Reporting should include time to detection, time to customer notification, loss recovery, agent involvement and repeat abuse of the same infrastructure.

Consumer protection also needs clear language. Users should know that no legitimate operator will ask for a PIN or one-time password, but warnings must be tied to real examples seen in the market. Generic safety messages lose force when campaigns become more local and more convincing.

Conclusion

The growth of African mobile money makes anti-fraud a piece of financial infrastructure. GSMA's figures show the economic scale; Red Card 2.0 shows the criminal scale. The right target is not only reimbursement after loss. It is reducing the time between detection, indicator sharing, blocking and user notification.

— ◆ —