Page 493 - ISC PROCEEDINGS 21.4

Transitioning to AI is not merely about installing software; it is a fundamental
transformation of the entire operational machinery onto a new technological foundation.
Legacy System Constraints: Traditional core banking systems often lack the flexibility
required to integrate with AI applications, which demand massive real-time data
processing capabilities. Specialist Shortages: There is a significant market deficit of
professionals who possess deep expertise in both finance and data science. Meanwhile,
traditional staff face the mounting pressure of displacement unless they undergo
comprehensive retraining.
Concentration and legal risks
Dependence on Third-Party AI Providers: Most banks currently lease AI
infrastructure from tech giants (Google, Microsoft, Amazon). If these partners experience
technical failures or outages, the global financial system could face simultaneous paralysis.
Incomplete Legal Frameworks: In Vietnam and many other nations, specific
regulations regarding legal liability when AI causes errors (e.g., resulting in the loss of
customer funds) are still under development. This creates a "gray zone" of risk for
banking institutions.
Table 2. Key Challenges for Banks in the Digital Era
Risk category Detailed description Impact
Deepfake attacks, algorithmic
Technical Loss of customer assets
vulnerabilities
Widespread service
Operational AI system failures or sudden outages
disruption
Reputational damage and
Ethical Algorithmic bias and discrimination
legal litigation
Difficulty in complying
Governance Lack of transparency in the AI "Black Box" with State Bank
regulations
Source: Synthesized by the author from academic research
2.3. Research methodology
To achieve the research objectives, this paper employs a qualitative research
method with an interdisciplinary approach (encompassing finance, technology, and
behavioral psychology). The research process is specifically executed through the
following three stages:
Data collection methods
Secondary Academic Data: The author synthesizes and analyzes the research works
of reputable international scholars, including Tony Boobier, Yves Hilpisch, and Pascal
Bornet, to establish a theoretical framework for Cognitive Banking, Invisible Banking, and
Intelligent Automation.
Empirical Data in Vietnam: Practical insights are gathered from annual reports,
financial statements, and development strategies (2020–2025) of five pioneering AI-
driven banks in Vietnam: TPBank, MB, Techcombank, VIB, and VPBank.
Case Study Research Method
This article employs the case study research method to conduct an in-depth analysis
of AI implementation within the target banks. The selection criteria for these five banks
are predicated on: (1) Pioneering Status: These institutions were the first to deploy eKYC,
RPA, or Cloud Banking within the Vietnamese market. (2) Demonstrated Efficiency:

492