Scam prevention depends on effective collaboration & AI
The impact of fraud and scams globally has reached over £1 trillion, highlighting the urgent need for effective scam prevention strategies. This staggering figure underscores the vast scale of mobile fraud, which affects individuals, businesses, and entire economies alike. The human costs of mobile fraud range from nuisance-level scams to life-changing events in which victims lose everything they have. There is also a disturbing trend towards blackmail through grotesque invasions of privacy, in some cases resulting in suicide.
The victims are not only those falling prey to the scams themselves, however. A far-reaching network of perpetrators also faces exploitation in the process, often under duress or coercion. For example, criminal syndicates in Southeast Asia enslave an estimated 200,000 people to commit mobile fraud on an industrial scale. This human cost of mobile fraud – and what can be done about it – was the focus of our first session at SEC CON 2025 at MWC this year, ‘Confronting Fraud: Securing a Safer Future’.
Accurate and rapid data sharing is crucial to scam prevention
As Toby Evans, Head of Economic Crime at the Australian Payments Network set out, strategic collaboration between key industries is essential to progress in this complex area. Australia has seen remarkable success in driving down digital fraud, with a 40% reduction in scam losses across industries, and one leading bank achieving a remarkable 70% reduction in fraudulent transactions.
Collaboration between financial institutions and the mobile industry was a game-changer here. ‘Do not originate’ lists have prevented number spoofing and dramatically reduced impersonation scams. This has helped prevent 2 billion fraudulent calls in a country of only 27 million people. Meanwhile, operators have educated financial institutions on a wide range of potential vulnerabilities. These include using SMS for one-time PINs and clickable links, mitigating mobile fraud via account takeover, as well as improving sender ID so users trust incoming communications enough to engage.
Toby stressed that accurate reporting and agreed taxonomies are essential. “Ideally, these should be globally agreed so we know we’re comparing like with like, and deduplicating data for maximum efficiency.” Achieving government buy-in has also been crucial, with the Australian government now imposing $50 million fines for companies failing to implement anti-fraud measures. This milestone act has set a precedent for global regulatory action and a framework for effective data sharing.
The problem will likely get worse before it gets better
So, what additional measures can we take to enhance scam prevention effectiveness? Recognising the scale of the problem is an essential starting point. 25% of the world’s population was scammed last year, according to the Global Anti-Scam Alliance (GASA). In some countries, fraud-related losses amount to nearly 3% of GDP. Scammers are particularly active in emerging markets like Brazil, India, and Southeast Asia, where digital literacy is still developing. Criminals exploit digital expansion, targeting those who are new to digital transactions.
The problem is nonetheless serious in developed markets. In the UK, fraud now accounts for 38% of all reported crime, with a growing reliance by perpetrators on digital platforms for financial crime. Criminals are using AI and deepfake technology to impersonate trusted individuals, in one case defrauding a prominent CEO of over $200,000 by using voice impersonation over a call.
“Fraudsters are becoming incredibly adept at manipulating victims,” explained GASA’s Managing Director Jorij Abraham, citing the rise of romance and investment scams. “One of the most alarming trends is scammers allowing victims to withdraw small amounts of money early on to gain trust, before convincing them to invest their life savings.” Such social engineering attacks are often based on disturbingly extensive datasets available to fraudsters. Scammers are also using such data to drive highly effective AI tooling. In the face of this increasing sophistication, AI-driven detection, industry-wide data sharing, and improved consumer education will be the foundation of future scam prevention.
Scam prevention means operators must fight AI with AI
AI is already playing a huge role in the fightback against these perpetrators. India’s Airtel, for instance, is processing 200 billion calls a day through AI checks against 26 different parameters. These include call velocity, device change frequency, the number of unanswered calls, disconnect speed and absences of outgoing messages. It has also developed effective algorithms to detect call spoofing and suspicious URLs in SMS. This allows Airtel to block around 1 billion attempts at mobile fraud each day with 99.7% accuracy, completely independent of user intervention or feedback.
AI, then, brings with it enormous potential for improvements in security by design. Some financial institutions, for instance, are now working with operators to integrate AI-driven risk scoring to protect digital transactions. “Every sector must contribute to scam prevention by adopting security-first principles,” said Toby Evans. “It is not enough to respond to fraud. We must prevent it from happening in the first place.” Mobile device and platform designers also have a crucial role to play. Google for instance has introduced OTP-blocking in Android phones when the user activates screensharing.
This was a common theme at the Summit – simply reacting to mobile fraud will never suffice. The focus must be on prevention and deterrence. Operators and their partners achieve this best through collaboration to make these crimes harder to commit, and more likely to result in arrest. Information sharing however remains a barrier, with a lack of common terminologies or approaches to collaboration.
Anti-fraud partnerships rely on shared taxonomies to overcome data siloes
“We still struggle with information silos,” noted Theresa Walsh, CIO at FS-ISAC. “Even within financial institutions, cybersecurity teams often do not communicate effectively with fraud prevention teams. That has to change.” And, as Cellcard’s CEO Simon Perkins illustrated with experience of the harsh reality in Cambodia, that inability to share data can have grave consequences. It can allow a slave facility to evade law enforcement, operate independent communications, or move operations over a border undetected.
Globe Telecom has set an excellent example here, through a memorandum of understanding on real-time data sharing with a major bank. The move has significantly improving response times and scam prevention rates. Globe has led successful initiatives to exclude clickable CTAs from P2P messages. They use machine learning algorithms to block scam keywords. The company also runs intensive customer awareness campaigns, targeting the elderly and very young.
Operators’ developers and cloud provider partners can draw on the GSMA Open Gateway to help accelerate solutions in this area. For the ecosystem at large however, the fight will rely on five key pillars. These include cooperation on data and technology. They implement multi-layered authentication and monitor emerging fraud trends. They also invest in AI tools and focus on user education. “Scammers are getting smarter, but so are we,” said Google’s Eugene Liderman. “With the right technologies, collaboration, and awareness, we can protect users and create a more secure mobile environment.” The fight against mobile fraud is ongoing and will not be over soon, but a more secure digital future is within reach.