The Number That Should Worry You Isn't 55%
INTERPOL's latest Africa cybercrime report puts a striking figure on something investigators have been describing anecdotally for years: AI is now linked to 55% of reported cybercrime across the continent. Online scams, business email compromise, romance scams, mobile money fraud, ransomware against infrastructure, deepfake-based sextortion, and synthetic identity fraud are all named in the report. INTERPOL's Cybercrime Director, Neal Jetton, summed up the mechanism plainly: "AI is automating every stage of a cyberattack, from reconnaissance and phishing to extortion and evasion."
The 55% figure is the one making headlines. The number I'd actually pay attention to is this one: losses more than doubled in a single year, from $192 million in 2024 to $484 million in 2025. That's not a story about a new category of crime appearing out of nowhere. It's a story about existing fraud — the same scam-center operations, the same mobile money schemes, the same romance scams — getting cheaper to run, faster to scale, and harder to catch, all at once.
AI Is a Force Multiplier, Not the Root Cause
It's tempting to read "AI-linked" as "AI-caused," and I think that framing lets the actual failure points off the hook. The report's own regional breakdown makes the underlying story clear: East Africa's exposure runs through mobile money fraud and ransomware, because mobile money is how a huge share of the region banks, and fraud follows money flow. Central and West Africa report scam centres in 72% of surveyed countries — a physical, organized-crime infrastructure that AI tools make more productive, not one AI invented. Southern Africa's high connectivity is drawing in threat actors from outside the region entirely, because better bandwidth means a bigger attack surface.
None of that is a story about artificial intelligence introducing a novel threat. It's a story about fraud operations that already existed getting a productivity upgrade, while the systems meant to catch them — cross-border law enforcement cooperation, mobile money KYC standards, telco-level fraud detection — mostly didn't get the same upgrade.
What Actually Needs to Move Faster
INTERPOL's own operational track record this year argues for where the leverage actually is: four coordinated operations led to over 1,500 arrests and roughly $100 million recovered. That's cross-border cooperation working at a scale that matches the problem, and it happened without banning or restricting any AI tool. Compare that to a losses trajectory that's still roughly doubling year over year, and the gap isn't a technology gap — it's a capacity and coordination gap. Mobile money providers, telcos, and regional financial regulators are the choke points where AI-enabled fraud actually gets stopped or doesn't, and that's an infrastructure and enforcement-funding question, not an "AI is out of control" question.
What It Means for You
If you're building or operating anything that touches payments, identity verification, or customer communication in African markets, the practical read here is that AI-enabled social engineering (voice cloning, deepfake video, hyper-personalized phishing) has to be treated as a baseline threat model now, not an edge case you'll get to later. That means investing in fraud detection that assumes synthetic audio and video are already in the attacker's toolkit, and treating KYC and transaction-monitoring gaps as urgent rather than routine backlog items. If your product operates across borders on the continent, INTERPOL's operational wins this year are a reminder that cross-border data-sharing and cooperation agreements are worth pursuing proactively — they're demonstrably where recoveries actually happen.
