In the age of AI, users are increasingly concerned about security risks and online fraud. Are digital banks leaving the elderly or low-income groups behind?
Dr Huy Pham: Trust is the foundation of digital finance – and once lost, it’s hard to rebuild. In the AI era, banks need to move from reactive to predictive trust-building. That means using AI not just for automation but for real-time risk detection, fraud prevention, and transparent communication with customers. In addition to issuing warnings and educational materials, banks need to provide comprehensive support for the victims of AI-related fraud and scams. Banks should help customers recover the losses after they have verified that their customers have been scammed.
At the same time, digital inclusion must be at the heart of transformation. Older and low-income users should not be left behind in this transition. Banks can introduce simplified interfaces, vernacular-language support, and micro-saving tools to bridge the gap. We must remember: the true success of digital banking is not measured by how fast it grows, but by how many people it brings along on that journey.
Security, data, and innovation: The three foundations of digital banking
The culture of innovation in the Vietnamese banking industry is still quite cautious. How can banks overcome “risk aversion” when testing new technologies?
Dr Fiza Qureshi: Banks in Vietnam tend to be cautious when adopting new technologies due to concerns over financial risk, reputational impact and regulatory requirements. However, avoiding innovation and experimentation can slow digital progress and limit the adoption of emerging solutions.
To address this, banks can adopt structured approaches such as selecting low-exposure use cases, collaborating with external partners, conducting small-scale pilots and aligning internal incentives to support responsible experimentation.
Using controlled testing environments, such as internal pilot zones or regulatory sandboxes, allows banks to trial new products with a small group of users, limited transaction values and clearly defined monitoring mechanisms. This creates space to innovate without compromising system stability. These controlled environments have been successfully applied in Singapore, Hong Kong, the UK, and Australia, demonstrating their effectiveness in reducing implementation risk.
Banks can begin with prototypes or pilot services under these controlled conditions, which support a test-and-learn approach. They can also co-develop solutions with technology vendors and fintech firms to share knowledge and implementation responsibilities. Starting with lower-risk applications, such as automating internal processes, improving customer onboarding or using AI for customer support, helps reduce direct financial exposure and lowers the pressure to achieve perfect outcomes immediately.
Another important aspect is cultural and incentive alignment. If staff are rewarded only for error avoidance, they will naturally avoid experimentation. Performance systems should recognise safe, well-documented testing and learning processes so that managers and employees feel supported when piloting new ideas. This shift encourages progress while maintaining accountability and control.
In terms of policy, what adjustments does Vietnam need to prioritise for sustainable development of digital banks, especially in data management, cybersecurity and user protection?
Dr Fiza Qureshi: For data management, Vietnam needs to prioritise and enforce national data-classification schemes and financial consumer protection standards. Banks need to follow clear guidelines under the new data law regarding handling of financial and credit information, privacy requirements and cross-border data transfers. The guidelines on open banking and open API should be aligned with the fintech sandbox to allow digital banks to integrate APIs safely. A clear pathway for virtual bank licensing with appropriate capital and customer-protection requirements would support market development. Banks should also conduct regular model validation for automated KYC and ongoing monitoring to address data quality and potential bias where AI is used.
On cyber security, the State Bank of Vietnam should require mandatory incident reporting and coordinated response frameworks, rather than relying on voluntary cooperation. Establishing a sector-wide security operations centre and real-time threat intelligence sharing for banks, e-wallets and payment firms would strengthen readiness. In addition, publishing frequent scam alerts and fraud typologies can support prevention efforts. For user protection, policies should reinforce strong customer authentication, step-up verification for first-time payees, clear warning prompts before high-risk transfers and rapid freezing of suspected mule accounts. Standardised refund and redress mechanisms for authorised push payment fraud should also be prioritised to enhance trust and safety.
Story: Nhu Quynh
This article was translated from Vietnamese with permission from CafeF. Read the original article here.
Thumbnail image: panuwat – stock.adobe.com | Masthead image: Tierney – stock.adobe.com
