AI-boosted mobile money fraud
SIM swap, voice deepfakes, automated social engineering: AI has changed the scale of mobile money fraud far faster than most defenses have scaled up in response. A tour of the landscape with real cases from West Africa.
October 17, 2026 · SunuCyberSecurity team
of reported cybercrimes in Africa involve AI (36 member countries)
in cybercrime losses in 2026 — more than double since 2024 ($192M)
of countries cite mobile money fraud as the most common scam type
of countries confirm organized scam centers operating on their territory
Source: INTERPOL, African Cyberthreat Assessment Report 2026
Mobile money has become the reference financial infrastructure for hundreds of millions of people in West Africa — and therefore a prime target. What has changed recently is the scale: generative AI automates three techniques that already existed separately, but that used to reach a limited number of victims at a time. Here is where the threat actually stands, backed by documented cases rather than projections.
1. AI-boosted SIM swap
AI can reconstruct a credible identity from real and fabricated data — enough to convince a mobile operator to transfer a number to a new SIM card. Once the number is hijacked, the OTP codes and linked mobile money accounts go with it. Three ingredients make the technique more dangerous than before: synthetic identities generated from real data leaks, automated bypassing of verification controls, and the chain hijacked number → intercepted OTP → emptied account, executed within minutes.
Kenya, 2025
+327%
in reported SIM swap cases — over 123,000 fraudulent SIM cards issued. East Africa is identified as the main zone for this technique.
Source: INTERPOL 2026 / Business Daily Africa
2. Voice deepfakes
A few seconds of public audio — a voice message, a video shared on social media — are now enough to clone a voice convincingly. The warning sign is clear: attempts are exploding far faster than detection tools. Pindrop's 2025 Voice Intelligence & Security Report found a +1,300% rise in synthetic-voice fraud attempts in 2024 — from about one attempt per month to seven per day — with +149% targeting banks specifically and +475% targeting insurers.
Hong Kong, 2024
An employee transfers $25.6M after a deepfake video call impersonating his CFO.
France, 2026
"Silent calls" use a voice cloned from recorded samples to ask the victim's contacts for money (OECD.AI Incidents Monitor).
Ghana, 2026
Experts warn: a few seconds of voice are enough to impersonate a mobile money agent or a relative over the phone.
The same principle applies to mobile money: a cloned voice impersonating a relative or an agent is enough to trigger a transfer.
Sources: Pindrop · business press · OECD.AI Incidents Monitor · MobileIDWorld
3. Automated social engineering
"Fraud-as-a-service" kits sell stolen data and verification-bypass tools ready to use — no technical skill required (Cifas, UK). AI-generated phishing and SMS reportedly reach, according to one vendor (Infobip, a figure to cross-check), up to a 54% click rate on AI-assisted smishing campaigns. Synthetic identities are used to open bank accounts, take out mobile micro-loans and register SIM cards. In Nigeria, phishing specifically targets BVNs, passwords and OTP codes (NIBSS).
Ghana, Q1 2025
$1.3M
lost to mobile money fraud in a single quarter (INTERPOL)
Continental estimate
$4Bn/yr
in estimated mobile money losses across Africa, SIM swap the leading cause (secondary source, to cross-check)
Sources: INTERPOL · Cifas · Infobip (vendor data) · Modern Ghana / MyJoyOnline · TechNext24 · NIBSS
Country watch — West Africa
Seven countries, seven documented cases — the technique is actively attempted everywhere, not just "elsewhere."
Senegal
In Guédiawaye, a Sonatel agent foiled a SIM swap attempt in August 2025: an individual with forged documents was trying to take control of SIM cards to access victims' Wave and Orange Money accounts. In June 2026, six foreign nationals were arrested in Ngor (Dakar) with telecom fraud equipment, including SIM box gateways — a sign of organized infrastructure, not isolated incidents.
Source: Osiris.sn
Burkina Faso
Orange Money Burkina Faso detected internal fraud in its distribution network in July 2026, with artificial transactions on the "Gold 1" merchant offer. In Bobo-Dioulasso, a fake withdrawal SMS shown to Orange Money counter agents was enough to obtain money with no verification (24 months in prison). A network of seven foreign nationals was also arrested after ~300 calls in 24 hours to identify Orange Money account holders — reconnaissance an AI system would make instant, at scale.
Source: Burkina24 · AllAfrica
Côte d'Ivoire
1,232 e-wallet fraud cases were recorded on the police platform in a single year, for about 55M FCFA in losses. At a public panel in Abidjan, a Wave executive and a Deloitte risk advisor confirmed that social engineering and SIM swap remain the two most frequent attack vectors — which is precisely what prompted the BCEAO instruction requiring, since 2021, real-time automated anti-fraud controls from all mobile money operators in the WAEMU zone.
Source: CIO Mag
Togo
The National Cybersecurity Agency (ANCY) estimates digital scam damage at 500M FCFA in 2025 — a figure the agency itself considers likely underestimated. The three methods it identified: identity theft, fake payment confirmation messages, and calls directly requesting a secret code or transaction validation.
Source: ANCY, via AllAfrica
Benin
A documented cross-border case: a fraudster operating from Nigeria used a SIM card fraudulently registered in Benin to empty a mobile money account... in Kenya. The African Union's Malabo Convention (2014), meant to frame judicial cooperation against this kind of continent-wide cross-border fraud, remains largely inoperative due to insufficient ratifications.
Source: FinancialAfrik
Mali
The BCEAO documents the "hypnosis scam": through conversation alone, the fraudster leads the victim to reveal their own Orange Money secret code, without ever hacking or manipulating a device. In Bamako, two brothers were arrested with a 32-port SIM box and 276 Orange chips at their home — the logistics of large-scale fraud.
Source: BCEAO · CIO Mag / Burkina24
Sierra Leone
On August 18–19, 2026, the government convened the central bank, the telecom regulator, operators, cybersecurity agencies and civil society to build a national roadmap against mobile money fraud. Minister of Communications Salima Bah acknowledges that digital financial fraud is becoming "increasingly sophisticated and widespread" — identity theft and fake WhatsApp accounts top the reports.
Source: Government of Sierra Leone
Warning signs to know
Early warning signs
—Sudden loss of network or inability to send SMS
—Operator notification of an unrequested SIM card change
—Unusual login failures, especially on a banking app
—Unsolicited call asking you to read out a code received by SMS
—A voice with a slightly robotic tone, unusual pacing or intonation
Protective habits
—Ask your operator for a dedicated PIN required for any SIM card change
—Prefer an authenticator app over receiving codes by SMS
—Contact your operator and bank immediately if network loss is unexplained
—Never share a code received by SMS, even with someone claiming to be official
—If a voice feels off, hang up and call back on a number you already know
Sources: ENISA · bank security resources (NatWest, Ulster Bank)
What every user should remember
1
No OTP code is ever shared, even with an "agent" on the phone — an operator or bank never asks for it
2
A familiar voice is no longer enough to verify identity — if in doubt, hang up and call back on a known number
3
A dedicated PIN, requested from the operator, protects against any unauthorized SIM card change
4
An unexplained loss of network is an urgent signal — contact the operator and bank immediately, not the next day
5
Urgency and emotion are social engineering's number one tool — a pressing request should slow the decision down, not speed it up
AI has changed the scale of the threat. Individual vigilance remains the first line of defense.
SunuProbe — SS7/USSD detection
Detect the telecom vulnerabilities exploited for SIM swap and mobile money fraud, at the SS7/USSD level of your own network.
Sources
INTERPOL (African Cyberthreat Assessment Report 2026) · Pindrop (2025 Voice Intelligence & Security Report) · Cifas · Infobip (vendor data) · NIBSS · Modern Ghana · TechNext24 · OECD.AI Incidents Monitor · Osiris.sn · Burkina24 · AllAfrica · CIO Mag · FinancialAfrik · BCEAO · Government of Sierra Leone · ENISA · SunuBarometer (sunucybersecurity.com/en/africa-cybersecurity-barometer). External figures not published by SunuCyberSecurity unless stated otherwise.
Put this into practice
A free scan, in a few minutes, no credit card.
Free scan