AI-Powered Fraud Arms Race — Criminals Outpace Defenders Across Africa’s Fintech Sector

NAIROBI, Kenya — Across Africa’s rapidly digitizing financial landscape, a silent and asymmetrical war is being waged. While fintech companies and banks deploy increasingly sophisticated artificial intelligence to detect and block fraudulent transactions, a new report warns that criminals are now using the same technologies — deepfakes, generative AI, and synthetic identities — to stay one step ahead. The Visa Mid-year 2026 Biannual Threats Report shows that from July to December 2025 alone, nearly $1 billion was linked to scam-related activity worldwide, making scams the single largest category of consumer payment fraud. Criminals are increasingly moving away from technical system breaches, instead exploiting human trust through AI-powered impersonation, urgency tactics, and highly convincing fake communications.

The rise of generative AI has fundamentally lowered the barrier to entry for fraudsters, enabling them to generate fake identities, manipulate images and documents, and create synthetic voices used in sophisticated attacks targeting financial platforms. According to a study cited by the Ecofin Agency, these tools allow criminals to carry out more complex and difficult-to-detect financial crimes at scale. In Nigeria, the region’s largest fintech market, the threat has become so severe that approximately 87.5 percent of fintech companies now use AI systems primarily for fraud detection — a clear indicator of the scale of the challenge.

The consequences are already visible in real time. According to findings from the Sumsub Identity Fraud Report 2025–2026, the share of sophisticated multi-step fraud attacks grew by 180 percent globally year-on-year, reaching 28 percent of all detected fraud. Nigeria also recorded the world’s highest share of synthetic document fraud in 2025, at 8 percent. The data reveals a critical vulnerability: as financial systems move toward instant, real-time payments, many institutions still rely on siloed onboarding systems, fragmented anti-money laundering tools, and manual compliance processes that operate at "human speed" while fraud operates at machine speed.

Africa’s overall identity fraud rate declined only marginally to 3.42 percent in early 2025, but the nature of the threat has shifted dramatically. While traditional document forgery has decreased — plummeting by roughly 80 percent year-on-year in Nigeria — synthetic identity document fraud has increased considerably in several key markets. "Enhanced verification tools have decimated traditional document forgery, but criminals are adapting with synthetic IDs and AI-powered scams," said Hannes Bezuidenhout, Sumsub’s VP of sales for Africa. The fraud prevention battle has become a race between innovation and adaptation.

For mobile money users, the stakes are particularly high. Cybersecurity firm Serianu estimates that Kenya alone has suffered cyber-related losses of roughly $230 million this year, with continent-wide losses nearing $5 billion. Payment fraud remains the most frequently reported security incident, with online and email fraud together accounting for 40 percent of all reported incidents and approximately 32 percent of total estimated losses. The report notes a growing trend of AI-enabled blended attacks that combine phishing, credential theft, ransomware, and impersonation across financial systems.

Many mobile-money services have failed to keep pace with the growing confidence users place in them, still relying on simple four-to-six-digit PINs instead of stronger, multi-layered authentication. "Trust in mobile money doesn't happen instantly — it's built one safe transaction at a time," warned cybersecurity engineer Allan Juma. He explained that when a user gets scammed, they often stop using mobile money entirely and warn friends and family to do the same, "creating a ripple effect that stalls progress for everyone". As William Makatiani, Serianu’s CEO, stressed, artificial intelligence, once heralded as a key safeguard, now requires defenders to fundamentally rethink how they protect systems and data.