•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
At the Digital Trust in Finance 2026 forum on May 12, MoMo co-founder and CEO Nguyen Manh Tuong opened his remarks with a personal account of an online fraud scam in which an elderly woman was tricked into transferring 20 million dong after receiving an urgent request from a Facebook account believed to belong to a close friend. After discovering the scam, the victim suffered for days, and her health later deteriorated after being hit by a car. Tuong said the incident reflects not only the financial loss but also the broader disruption to family stability and peace of mind.
Tuong cited data from Vietnam’s Ministry of Public Security showing that last year Vietnamese people lost 8,000 billion VND to online fraud. He also referenced GASA data indicating that two-thirds of victims experience prolonged stress, while nearly half report mental health impacts afterward. He stressed that the harm is not limited to older people: “On our platform, 50% of fraud victims are students under 22,” he said, adding that scammers are targeting both money and trust, including among younger users.
Tuong said fraud is no longer limited to isolated individuals sending scam messages from a keyboard. He pointed to multiple indicators of how attacks are scaling and accelerating. According to an Industry Report, worldwide deepfake voice attacks increased by 2,100% in 2024. He also cited Sift data that an attack occurs every five minutes, and that 82% of emails are AI-generated—potentially rising to 100% at the current pace. He added that fraud schemes are spreading 40% faster than before.
“Not only the numbers are scary; more and more actors can participate, without technical know-how or large capital. Any bad actor can start a large-scale attack campaign very quickly,” Tuong said.
Tuong described an asymmetry between attackers and defenders. He said fraud has evolved into an organized industry with cross-border division of labor, supported by technology that targets psychology and emotion. By contrast, he said organizations often need days or even months to process an attack. He also highlighted cost asymmetry driven by AI.
On the attacker side, Tuong said it can cost as little as $20 per month to run a fraud campaign while yielding over $100,000, with higher profits for more complex campaigns. He added that, based on statistics from Deloitte, Chainalysis, and Sift, AI-assisted fraud profits are 4.5 times higher than traditional methods.
He noted that Asian banks and financial institutions are increasing investment in AI to combat fraud. According to Chainalysis, the investment share rose from 16% to 68%—more than fourfold in two years. However, he said costs are rising while user trust continues to fall: up to 76% of users are willing to leave a platform after being scammed, creating customer losses and additional hidden costs.
Tuong said a different anti-fraud approach is needed. He argued that traditional thinking focuses on detecting a transaction and blocking it, but many successful scams occur because users complete transactions after being psychologically manipulated.
MoMo’s leadership said the strategy shifts from “block the bad guys”—which is difficult to achieve fully—to “walk with the good users,” helping them avoid wrong decisions. The focus is not only on blocking transactions, but also on detecting anomalous behavior, understanding context, and intervening at the right moment. Tuong emphasized that “context” is particularly important, noting that criminals often try to isolate victims emotionally.
On the role of AI, Tuong said that criminals using AI does not mean companies only need stronger AI models. He stated that “90% of the problem isn’t in the model or algorithm,” arguing that the critical value is using signals from algorithms to influence user behavior at the right time. If interventions are not timely, users may still ignore warnings and proceed with transactions.
MoMo reported that more than 10,000 platform users were blocked before any damage occurred, which Tuong said is equivalent to about 15 trillion VND per year. He also noted that each MoMo transaction is processed in 100–300 milliseconds. According to MoMo’s early results, for every 1,000 users who receive a system alert, 995 stop the transaction.
Tuong said the platform’s approach helps avoid about 44 billion VND in potential fraud daily. He also cited 29,000 anomalous transactions alerted and more than 10,000 users blocked before damage occurs, equivalent to about 15 trillion VND per year.
Despite these early results, Tuong acknowledged that the platform still misses many victims. He said the fraud journey often begins earlier than the point where MoMo can intervene: attackers identify targets, contact them via social media, messaging apps, or calls, build trust, and only then request a transfer.
“When a victim transfers money on MoMo or a banking app, we are effectively watching the 89th minute of a 90-minute movie,” he said. Tuong added that at that stage the system has only about five seconds to block an online transaction, after which time is counted in minutes. He said that in less than 30 minutes, money may be drained, while investigations can take days.
He attributed the challenge not only to speed, but also to fragmented data and signals across companies and sectors, and to fragmented coordination. He said attacks are proactive and continuous, with each side seeing only part of the picture. “We are still fighting quite unevenly,” Tuong concluded.
Premium gym chains are entering a “golden era” that is ending or already in decline, as rising operating costs collide with shifting consumer preferences toward more flexible, community-based ways to exercise. Long-term memberships are shrinking, margins are pressured by higher rents and facility expenses, and competition from smaller, more personalized…
