Where Does Banking Expertise Persist in the AI Era?

     Banks are rapidly advancing toward full digitalization. The integration of regulatory frameworks and operational guidelines into computerized systems aims to standardize procedures, minimize human error, and mitigate legal and reputational risk. In principle, this objective is well justified. Digital infrastructures reduce the likelihood of omission and error by enforcing procedural compliance.
     However, as technical accuracy improves, a parallel transformation emerges within frontline operations. As employees become increasingly dependent on system-guided workflows, operational efficiency improves, yet critical reflection on procedural necessity correspondingly declines. Questions such as “Why is this process required?”, “What specific risks is it designed to prevent?”, and “Where do its exceptions and boundary conditions lie?” gradually recede. When such inquiry disappears, employees shift from being interpreters of institutional knowledge to executors of predefined digital routines. At this stage, customers may begin to question the professional authority of bank personnel.

     The expansion of AI further sharpens this flow. Customers increasingly utilize AI tools with greater fluency than bank employees to compare product structures and to interrogate interest rates, fees, taxation, and risk exposure. Information that was once monopolized by banking professionals is rapidly becoming democratized, and explanatory competence alone is no longer sufficient to sustain trust.
     This does not render financial professionals obsolete; rather, it redefines expertise, as authority can no longer rely on knowledge monopolization. Instead, they must demonstrate credibility through the governance of decision quality and accountability. Banking is not merely an information industry but fundamentally a trust-based industry, because trust is produced less by the correctness of an answer than by the transparency and integrity of the processes through which that answer is generated and implemented.
     This principle applies consistently across banking operations. Even in lending alone, the number of products is large and their conditions are complex. However, the “most suitable loan” is not determined by an interest rate table alone. The optimal structure varies depending on how contextual information is assessed—including income stability, existing debt structure, repayment priorities, and other contextual factors.
     The same logic applies to deposits. The decision is not determined by the highest interest rate alone. Maturity, liquidity, the likelihood of meeting preferential conditions, and the alignment of product duration with financial purpose all influence the choice.
     Ultimately, the quality of AI-generated conclusions depends on accurate input and on employees’ ability to interpret and select among multiple outputs. Professionals must therefore remain interpreters of context, not mere system operators.
     Responsibility becomes even more significant in investment products. Within the financial consumer protection framework, investment recommendations combine suitability assessment, risk disclosure, and formal documentation; they are fundamentally different from casual suggestions. The fact that a sentence or recommendation is generated by AI does not reduce professional responsibility. What matters is not simply what was recommended, but why the product is appropriate for that specific customer, how risks were explained and consent obtained, and whether the process remains verifiable after the fact. In this sense, financial professionals in the AI era should be understood not as people who let AI speak on their behalf, but as actors who evaluate AI outputs within regulatory rules, risk considerations, and customer context, and translate them into accountable advice.
     AI multiplies answers, but trust depends on the quality and explainability of the decision process. In the AI era, banks create value not by providing information, but by governing decisions responsibly. Professional expertise is therefore shifting from knowledge ownership to accountability and trust design. 

 Kim Won-ok

[Professor, school of Business Administration]