Vietnam Launches AI Alert System to Combat Evolving Online Fraud

Vietnam is establishing a collaborative alert system utilizing artificial intelligence (AI) to combat increasingly sophisticated online financial fraud. The initiative aims to connect signs of criminal activity scattered across various platforms, from social media to bank accounts, to detect risks before victims are harmed.


On November 13, the National Cybersecurity Association (NCA) of Vietnam held a seminar in Hanoi to discuss cooperation between government agencies and private companies in sharing information about suspected criminal activities, according to local media including the Ministry of Public Security and VnExpress.


Data released by the Ministry of Public Security indicates that over 1,800 online fraud cases were reported nationwide from January to September this year. Although this marks a decrease of more than 20% compared to the same period last year, the methods employed by criminals, particularly those involving AI for identity theft and cross-border operations, are becoming more complex.


Nguyen Huy Luc, Deputy Director of the Cybersecurity Department (A05) of the Ministry of Public Security, noted, "Criminal organizations are approaching victims using overseas messaging apps and Voice over Internet Protocol (VoIP). They are particularly using AI to manipulate images and voices to gain trust before stealing money through virtual assets."


Connecting Fraud Indicators Across Platforms


The Vietnamese government's initiative to establish a collaborative alert system stems from the challenge that online fraud spans multiple services while related information is dispersed among individual companies.


According to MoMo, a Vietnamese e-wallet service provider, fraudsters typically contact victims via social media, then continue the conversation through messaging apps or phone calls to induce money transfers. However, financial institutions often struggle to identify how victims were deceived, even if they detect suspicious transactions at the final transfer stage.


In response, Vietnamese authorities are promoting an online risk alert network involving banks, e-wallet providers, telecommunications companies, social media operators, e-commerce, and logistics firms.


The proposed alert system would involve each institution sending suspicious phone numbers, accounts, internet addresses, and social media accounts to a central processing system. AI would analyze the information collected from various agencies to assess risk levels and issue warnings to relevant organizations.


For instance, if a particular phone number receives multiple fraud reports, the system would analyze associated social media accounts or financial accounts to identify potential additional risks early. Institutions receiving alerts would then determine necessary actions within their domains and share the outcomes.


Similar information-sharing systems are already in operation abroad. The Global Signal Exchange (GSE), an online fraud information-sharing platform, reported that 130 organizations participated in a seminar, collecting 1.7 billion risk signals over 22 months. Examples from Singapore, the UK, and Belgium highlighted successful identification and response to malicious accounts, websites, and advertisements through inter-agency information sharing.


Challenges of Expanding Information Sharing While Protecting Privacy


However, experts have pointed out that for the collaborative alert system to function effectively, it must ensure not only information sharing among companies but also the protection of personal data and the accuracy of alerts. Nguyen Hong Quan, Deputy Director of the Cybersecurity Department, emphasized during the seminar, "A fragmented and isolated alert system in a connected cyberspace cannot prevent or respond to fraud effectively."


He outlined the need for a practical information-sharing framework, a reliable data security foundation, and ongoing participation from businesses to establish the alert system. He cautioned that without clearly defining the purpose, scope, and management responsibilities of shared information, issues of privacy violations and reluctance from companies to provide information could arise.


Filtering out false alerts is also crucial, as merely classifying a specific phone number or account as suspicious does not necessarily imply involvement in actual criminal activity. Therefore, verifying the accuracy of information and correcting false alerts are essential conditions for the operation of the alert system.


Initial personal responses are also being emphasized. Earlier, on August 14, Ngo Minh Hieu, an expert from the National Cybersecurity Association, advised, "If online fraud is suspected, disconnect from the internet and contact the bank to block the account within the first 15 minutes." He also stressed the importance of changing passwords and preserving evidence such as text messages for reporting.


As the scope of information sharing expands, addressing the risks of personal data breaches and damages from false alerts remains a critical challenge. Establishing operational standards that ensure the swift detection of criminal signs while maintaining the accuracy and security of shared information is expected to be a key issue moving forward.





* This article has been translated by AI.