As synthetic intelligence (AI) turns into extra deeply embedded in digital communications, telecommunications fraud is evolving at an alarming tempo, costing operators billions globally and threatening buyer belief throughout markets.
In 2025, the intersection of AI instruments, social engineering, and more and more subtle assault vectors has created an ideal alternative for fraudsters and a strategic crucial for operators.
Past direct monetary losses, telecom fraud threatens buyer loyalty, regulatory standing, and public belief. In a hyper-connected and digitised economic system, a breach in telecom safety can have cascading results throughout fintech, banking, and even public infrastructure.
Learn additionally: 16,000 telecom sites at risk as diesel supply crisis deepens
“Fraud is not a back-end subject, it’s a boardroom dialog. Operators who act now will defend their revenues and reputations. Those that delay might not get well,” Harsha Angeri, vp, Company Technique and head of AI enterprise at Subex, asserted.
Angeri warned that telecom firms should urgently rethink their fraud detection fashions.
“Right now’s fraud panorama is dynamic. We’re seeing new types of exploitation that bypass legacy techniques with ease. Operators must proactively determine traits, share intelligence, and implement adaptive techniques to remain forward,” Angeri stated in a current professional article.
In line with Angeri, the highest 10 telecom fraud traits telcos can not afford to disregard this 12 months embrace the AI-powered scams and generative AI exploits
Fraudsters are actually leveraging instruments like generative AI to imitate communication kinds, automate rip-off interactions, and engineer faux however extremely persuasive customer-facing messages, making conventional fraud detection out of date.
One other one is immediate injection assaults on AI fashions. A rising cybersecurity concern entails manipulating generative AI prompts to bypass safety filters or extract delicate system information. This presents a brand new class of fraud, particularly in customer support AI bots or digital brokers utilized by telecom corporations.
Different fraudsters’ techniques are voice cloning and deepfake audio calls. Utilizing AI, scammers can now convincingly clone the voices of telecom executives or monetary controllers to authorise wire transfers or acquire entry to safe platforms. One misstep might result in multi-million-naira losses.
There may be additionally the bypass draud (SIM field fraud). This method, which reroutes worldwide calls to appear to be native site visitors and dodge tariffs, stays prevalent, costing African operators alone a whole bunch of hundreds of thousands yearly. Fraudsters are refining it with AI automation and quicker routing instruments.
Additionally, SIM swapping and account takeovers have change into main threats. Fraudsters more and more use Telegram and different darkish internet platforms to impersonate telecom suppliers, harvesting private information and porting numbers to realize entry to one-time passwords (OTPs), financial institution accounts, and crypto wallets.
Furthermore, flash calls and OTP hijacking are a part of the threats. Flash name verification, the place a system calls a consumer’s cellphone quantity to confirm id, is being exploited utilizing spoofed caller IDs. This permits scammers to bypass two-factor authentication (2FA) checks undetected.
The Wangiri fraud (one-ring callbacks) can also be common. AI-powered techniques are actually automating Wangiri scams, the place victims obtain a missed name from a premium-rate worldwide quantity and are charged exorbitant charges after they name again.
Inner income share fraud additionally happens, the place fraudsters generate synthetic site visitors to premium numbers, sharing the fraudulent income with complicit carriers in unregulated markets, through the use of hacked PBX techniques and manipulated VoIP gateways.
Learn additionally: Nigeria records 1,700 weekly attacks on telecom infrastructure
For the SIM closure and id fraud, criminals trick prospects into deactivating their SIMs or revealing private information, by posing as regulators. The harvested info is then used for id theft or large-scale monetary scams.
The AI-driven smishing and phishing campaigns are additionally threats. Extremely focused SMS and electronic mail campaigns, usually generated or enhanced by AI, are convincing sufficient to trick even tech-savvy customers into divulging credentials. The usage of spoofed telecom portals compounds the chance.
Multi-pronged response, non-negotiable
A multi-pronged response is non-negotiable, as Angeri advisable, noting that operators should undertake a multi-layered technique that mixes conventional fraud administration techniques with AI-driven analytics, buyer training, and real-time intelligence sharing.
“Operators should combine behavioural biometrics, blockchain for name authentication, and stronger id verification mechanisms like device-based and biometric authentication,” he stated.
Moreover, regulatory compliance, cross-industry collaboration and proactive incident response plans are not elective as they’re foundational to fraud resilience, Angeri asserted.
In his response, Dinesh Balsingh, CEO of Airtel Nigeria, instructed BusinessDay that the telco is tackling spams and unsolicited promotional messages, which stay a persistent supply of buyer frustration and fraud danger.
Balsingh said the corporate is investing in superior instruments to detect and block spams, significantly bulk SMS content material despatched with out consumer consent.
Learn additionally: Active telecom subscriptions fall below 172m in June amid SIM-NIN cleanup
“We additionally do spam detection, as a result of we all know prospects are complaining that they obtain a whole lot of spam messages. We’re in a position to determine the place these messages are coming from and block them. If there’s a bulk SMS despatched with out subscription, we cease it,” Balsingh stated.
He famous that Airtel deploys analytics and buyer expertise APIs to hint the origin of spam site visitors and take preemptive motion, making certain prospects solely obtain related, consent-based communication.
