Economic Essay Contest Finance in the age of AI: Why trust matters more than efficiency
Have you ever felt frustrated after spending hours interacting with a banking chatbot, only to receive the same repetitive response: “Sorry, I do not understand your request”? This highlights a core tension in the digital era: While technology operates with remarkable speed, it often fails to truly understand human needs.
Currently, major financial centers are investing heavily in artificial intelligence (AI). Despite this investment, a gap remains between technological capability and user experience, as efficiency is often mistaken for service quality. To remain sustainable, financial institutions must evaluate both the benefits and risks of AI to ensure long-term growth.
Accelerating services or sacrificing empathy?
It is undeniable that AI has driven significant advancements in financial services. One clear illustration is MoMo in Vietnam, which utilizes AI to automatically categorize consumer spending and enable facial-recognition payments in only three seconds. Similarly, Bank of America's virtual assistant Erica has processed more than 1.5 billion interactions while proactively alerting customers when their spending behavior indicates a risk of financial overextension. These applications demonstrate that AI can significantly enhance processing speed, personalization and convenience in banking services.
On the flip side, challenges emerge when financial institutions blindly pursue “100 percent automation” as a cost-reduction strategy. Under such circumstances, operational efficiency may improve while service quality deteriorates. Many customers in Korea have expressed frustration after being trapped for up to 40 minutes in repetitive chatbot conversations without resolving their issues. Public criticism intensified further when KB Kookmin Bank faced backlash for its plan to replace 240 call center employees with AI systems.
Instead of fully replacing human workers, the future of financial services depends on how effectively institutions combine AI with human judgment. While AI excels at speed and pattern recognition, it lacks the ability to interpret nuance, especially in situations involving stress, confusion or financial vulnerability. When customers are forced to navigate rigid automated systems without the option of human support, frustration quickly turns into distrust. Over time, this erosion of trust can become more costly than any short-term savings gained through automation. A more sustainable approach is selective integration: AI should handle routine tasks such as basic inquiries and data processing, while human advisers step in when problems require empathy, flexibility or critical thinking. In this model, technology enhances service efficiency without undermining the relational foundation that financial institutions depend on. Ultimately, while AI can simulate interaction, it cannot duplicate the genuine empathy that forms the bedrock of financial trust.
Expanding access or deepening the transparency gap?
In lending markets, AI is increasingly expanding access to credit for groups traditionally disadvantaged by conventional evaluation models. Instead of relying solely on credit scores, platforms such as Upstart in the United States analyze approximately 2,500 variables. As a result, loan approval rates for Black and Hispanic applicants have increased by 35 to 46 percent. Similarly, digital banks in Korea are leveraging AI to enhance financing accessibility for freelance workers.
That said, a major limitation of AI lies in its “algorithmic black box” nature, meaning that systems can produce decisions while users struggle to understand the rationale behind them. In practice, if a customer is denied credit approval and receives only a vague response, such as “Your application was declined due to internal scoring parameters,” this ambiguity can easily create perceptions of unfairness, especially when users are unable to challenge or improve future outcomes due to a lack of clear feedback.
Therefore, a necessary solution is the implementation of explainable AI (XAI). Instead of merely delivering final outcomes, the system should clearly identify the factors influencing each decision. For instance, customers could receive a specific explanation such as, “Your loan application was not approved because your income has been unstable over the past three months, and your current credit card outstanding balance remains high.” Such transparent, data-based communication not only helps customers understand the decision-making process but also builds trust in technological systems.
The race between protection and fraud
In financial security, AI is becoming a strategic digital shield. Mastercard analyzes one trillion data points in milliseconds to block fraud, while Korea’s ASAP system connects 130 organizations for real-time information sharing. Vietnam’s banks also use AI to detect unusual transactions from unfamiliar devices.
Conversely, cybercriminals are simultaneously exploiting AI to develop more sophisticated scams.
Deepfake technology enables fraudsters to impersonate voices or identities, making scams more convincing. In Vietnam, cases involving impersonation of relatives or authorities have become increasingly common. As these schemes grow more realistic, users become more vulnerable. At the same time, biased training data may lead to incorrect risk assessments, potentially disadvantaging legitimate customers.
In light of these challenges, effective solutions must go beyond technological investment and focus on governance mechanisms. First, a “cooling-off” mechanism can be implemented for high-risk transactions. When unusual behavior is detected, transactions are temporarily delayed, allowing users to cancel if fraud is suspected. Second, a “human-in-the-loop” approach ensures that high-risk cases, such as large transfers, rapid transactions or high-value loan decisions, are reviewed by human experts before final approval. Third, context-based multi-factor authentication should be strengthened by combining biometrics, device recognition and location tracking. Fourth, real time fraud alerts via mobile applications can help users identify common scams, including impersonation tactics.
While the influence of AI is pervasive, the future competitive advantage of financial institutions will not be determined by who possesses the most powerful computational systems, but by those who apply technology in the most human-centered manner. Leading institutions will be those that transform AI complexity into simple and empathetic customer experiences, expand access to credit while preserving transparency in decision-making, and place technology within strict ethical boundaries to protect people.
In other words, technological innovation is an inevitable requirement for development. Nevertheless, if that innovation gradually diminishes the human element, then it can no longer be regarded as genuine progress but rather a step backward in systems that are meant to serve people.