Economic Essay Contest The future of AI in finance: Building trust before fraud happens
A few months ago, I received a call from an unknown number, and the caller identified himself as a government agency in Korea. As an international student, I often deal with unfamiliar bureaucracy and calls, so everything seemed quite convincing. The person on the phone spoke confidently, mentioned my personal information and insisted that I take immediate actions.
For a second, I almost believed him.
But suddenly I received a warning from my banking app about voice-phishing scams. This warning was handled by an artificial intelligence (AI) assistant manager, not a bank staff member. The warning landed me to immediately end the call. Later, I learned that similar scams are common across the country.
What stayed with me was not the scam itself, but the realization that I trusted a warning generated by a financial system more than the voice of the person speaking on the phone.
That experience changed the way I think about AI in finance. After this call, I thought about the future of banking. I believe that banking should be a system that people can really trust, especially when things are uncertain. More importantly, it is about understanding that different clients face different risks and therefore require different forms of protection.
AI is already changing the way financial institutions serve their customers. AI assistants can answer questions instantly, provide 24 hours support and help banks respond more efficiently to customer needs. However, customer service should not become fully automated. Looking back on that phone call, I realize that it wasn’t just the technology itself that helped me, but also the feeling that someone or something cared about my safety.
As AI becomes more common, financial institutions should develop hybrid service models that combine technological efficiency with human support. However, customer service should not follow a one-size-fits-all approach. An elderly customer may need a simplified interface and direct assistance, while an international student may require multilingual guidance. AI should help banks identify these different customer mental models and provide support accordingly. Customers need quick answers, but when making important decisions, they also need human support, because people are not yet ready to have full trust in AI.
My experience was only one small example of a much larger problem. Fraud has become so widespread that businesses around the world lose an estimated 5 percent of their annual revenue to fraudulent activity, while global cybercrime damages are expected to exceed $10.5 trillion annually. In an environment where millions of transactions occur every second, financial institutions can no longer rely on human monitoring alone. At the same time, AI should not completely replace human judgment. Instead, it should complement human expertise by identifying suspicious patterns at a scale that people cannot manage on their own. In this way, AI becomes valuable as a tool for prevention.
The warning that I received earlier from the banking system came to me as part of this overall change. The change is that banks are becoming less concerned with reacting to fraud and more with predicting and preventing fraud. The Korean ASAP program is an example of this trend in that it assists financial institutions in sharing information about fraud cases. With financial crimes becoming ever more international, systems like this might prove valuable for other banks as well.
Nevertheless, this only covers half of the battle. In addition to predicting financial fraud, financial institutions need to realize that different individuals face different kinds of financial risk.
This is why I think the next phase of advancement in AI should involve personalized security. Banks already use AI to personalize financial products; they should also use it to personalize security. An elderly customer living alone, a first-time digital banking user and an international student may all have completely different security concerns. Instead of offering the same security service for everyone, AI can help recognize high-risk situations and protect accordingly. Customers might receive stronger fraud warnings, simplified explanations, additional verification procedures or immediate access to human assistance when unusual activity is detected. In the long run, the most successful financial institutions will not be those that offer the same protection to every customer, but those that understand which customers need protection the most.
The phone call lasted only a few minutes, but it changed how I view the future of finance. AI can improve customer service, create more personalized financial products, and strengthen financial security that is unachievable if done by just human beings. But, the success of such technology does not just depend on efficiency or accuracy, but on the degree to which individuals trust it.