As artificial intelligence (AI) becomes more deeply embedded in digital communications, telecommunications fraud is evolving at an alarming pace, costing operators billions globally and threatening customer trust across markets.

In 2025, the intersection of AI tools, social engineering, and increasingly sophisticated attack vectors has created a perfect opportunity for fraudsters and a strategic imperative for operators.

Beyond direct financial losses, telecom fraud threatens customer loyalty, regulatory standing, and public trust. In a hyper-connected and digitised economy, a breach in telecom security can have cascading effects across fintech, banking, and even public infrastructure.

“Fraud is no longer a back-end issue, it is a boardroom conversation. Operators who act now will protect their revenues and reputations. Those who delay may not recover,” Harsha Angeri, vice president, Corporate Strategy and head of AI business at Subex, asserted.

Angeri warned that telecom companies must urgently rethink their fraud detection models.

“Today’s fraud landscape is dynamic. We are seeing new forms of exploitation that bypass legacy systems with ease. Operators need to proactively identify trends, share intelligence, and implement adaptive systems to stay ahead,” Angeri said in a recent expert article.

According to Angeri, the top 10 telecom fraud trends telcos cannot afford to ignore this year include the AI-powered scams and generative AI exploits

Fraudsters are now leveraging tools like generative AI to mimic communication styles, automate scam interactions, and engineer fake but highly persuasive customer-facing messages, making traditional fraud detection obsolete.

Another one is prompt injection attacks on AI models. A rising cybersecurity concern involves manipulating generative AI prompts to bypass security filters or extract sensitive system data. This presents a new class of fraud, especially in customer service AI bots or virtual agents used by telecom firms.

Other fraudsters’ tactics are voice cloning and deepfake audio calls. Using AI, scammers can now convincingly clone the voices of telecom executives or financial controllers to authorise wire transfers or gain access to secure platforms. One misstep could lead to multi-million-naira losses.

There is also the bypass draud (SIM box fraud). This technique, which reroutes international calls to look like local traffic and dodge tariffs, remains prevalent, costing African operators alone hundreds of millions annually. Fraudsters are refining it with AI automation and faster routing tools.

Also, SIM swapping and account takeovers have become major threats. Fraudsters increasingly use Telegram and other dark web platforms to impersonate telecom providers, harvesting personal data and porting numbers to gain access to one-time passwords (OTPs), bank accounts, and crypto wallets.

Moreover, flash calls and OTP hijacking are part of the threats. Flash call verification, where a system calls a user’s phone number to verify identity, is being exploited using spoofed caller IDs. This allows scammers to bypass two-factor authentication (2FA) checks undetected.

The Wangiri fraud (one-ring callbacks) is also regular. AI-powered systems are now automating Wangiri scams, where victims receive a missed call from a premium-rate international number and are charged exorbitant fees when they call back.

Internal revenue share fraud also occurs, where fraudsters generate artificial traffic to premium numbers, sharing the fraudulent revenue with complicit carriers in unregulated markets, by using hacked PBX systems and manipulated VoIP gateways.

For the SIM closure and identity fraud, criminals trick customers into deactivating their SIMs or revealing personal data, by posing as regulators. The harvested information is then used for identity theft or large-scale financial scams.

The AI-driven smishing and phishing campaigns are also threats. Highly targeted SMS and email campaigns, often generated or enhanced by AI, are convincing enough to trick even tech-savvy users into divulging credentials. The use of spoofed telecom portals compounds the risk.

Multi-pronged response, non-negotiable

A multi-pronged response is non-negotiable, as Angeri recommended, noting that operators must adopt a multi-layered strategy that combines traditional fraud management systems with AI-driven analytics, customer education, and real-time intelligence sharing.

“Operators must integrate behavioural biometrics, blockchain for call authentication, and stronger identity verification mechanisms like device-based and biometric authentication,” he said.

Furthermore, regulatory compliance, cross-industry collaboration and proactive incident response plans are no longer optional as they are foundational to fraud resilience, Angeri asserted.

In his reaction, Dinesh Balsingh, CEO of Airtel Nigeria, told BusinessDay that the telco is tackling spams and unsolicited promotional messages, which remain a persistent source of customer frustration and fraud risk.

Balsingh stated the company is investing in advanced tools to detect and block spams, particularly bulk SMS content sent without user consent.

We also do spam detection, because we know customers are complaining that they receive a lot of spam messages. We are able to identify where these messages are coming from and block them. If there is a bulk SMS sent without subscription, we stop it,” Balsingh said.

He noted that Airtel deploys analytics and customer experience APIs to trace the origin of spam traffic and take preemptive action, ensuring customers only receive relevant, consent-based communication. (BusinessDay)