[Economy Pulse] AI Financial Consumer Protection: What Matters More Than Algorithmic Judgment

Identifying Unresolved Issues After "Case Closed"

Continuous Review and Work Improvement Needed

The "Financial Consumer Protection Performance National Report Conference" held on September 17 served as a showcase for how financial oversight has evolved over the past year. Among various achievements—such as organizational restructuring, greater activation of dispute mediation, and enhanced proactive oversight throughout financial product life cycles—what stood out the most was the use of artificial intelligence (AI). The Financial Supervisory Service is preparing a generative AI-powered civil complaint and dispute resolution portal that assists with complaint summaries and draft replies, and has already established an AI-based monitoring system for the virtual asset market. Financial companies, for their part, have either adopted or are in the process of adopting AI for product reviews and sales explanations, as well as complaint analysis and on-site response. Both supervisors and financial firms have chosen AI as a tool to boost their capacity to protect financial consumers. What now needs to be examined is how much consumer inconvenience and damage have actually been resolved through the use of AI.

For example, suppose a remittance was blocked due to suspected voice phishing, but after verification, it was confirmed to be a legitimate transaction and the restriction was lifted. While the financial company may record this action as complete, the consumer—unable to pay their rent on time during this period—still suffers consequences. Likewise, if a consumer receives a response to their complaint but gives up on follow-up inquiries because they cannot understand it, their issue remains unresolved. Just because a financial company marks the complaint as 'addressed' does not mean the consumer's problem has been solved. These types of disadvantages and inconveniences are difficult to capture by simply counting the number of cases processed.

The solution is to search for cases where problems are likely to remain even after being recorded as 'completed'. Typical examples include situations where a consumer interrupts the conversation during an AI consultation; reaches out again about the same issue; experiences a delayed lifting of remittance restrictions despite confirmation of a legitimate transaction; or raises an objection after receiving a response. Financial companies need to recognize these situations as signals indicating potential unresolved cases and review them separately. Even among 'completed' cases without such signals, a sample should be randomly selected for review in order to detect issues that may have been missed because consumers gave up on further inquiry out of exhaustion.

For the cases selected as such for additional checking, companies need to directly contact the consumers to confirm whether the issue has in fact been resolved. The questions can be simple and specific, such as: "Did you understand the guidance?" "Was your transaction or complaint resolved?" "Do you still need help from a representative?" If the answer is that their issue was not resolved, the details and submitted documents from the initial consultation should be passed on to a representative for continued handling. This lessens the burden on consumers to explain their situation from scratch and ensures the outcomes of any subsequent actions are confirmed.

It is the role of the Financial Supervisory Service to ensure that this kind of follow-up verification becomes a regular part of financial companies' daily operations. During audits and evaluations, it is important to sample ‘completed’ cases, review consultation records and follow-up actions, and closely monitor cases where repeated inquiries or prolonged transaction restrictions are common. If the same issues recur repeatedly, financial companies should be required to identify the underlying causes, appoint responsible persons for improvement, and specify a deadline for resolution, verifying afterward whether these measures reduce the occurrence of similar problems. Oversight that narrows the gap between a record of ‘completion’ and true problem-solving is needed.

Financial companies must ensure these reviews actually lead to improvements in their work processes. For example, the consumer protection department should identify unresolved cases every month and, together with the product, sales, and digital departments, analyze the causes and set out specific corrective measures and timelines. They should then check if the frequency of consultation abandonment and repeat inquiries for similar types of issues has decreased. The discovery of problems, correction, and tracking of effectiveness must become routine components of daily work.

To enable consumers to participate in this process, companies must clearly inform them of how to request further resolution. Instructions such as 'Unresolved Alert', 'Human Consultation', and 'Submit Additional Documents/Request Re-evaluation' should be displayed prominently in AI consultation windows and complaint replies. When a transaction is blocked, the process and expected time required for lifting the restriction must be clearly stated. Consumers, in turn, should confirm whether the response matches their actual situation and be proactive in requesting re-evaluation by providing any missing facts and supporting documents. Financial consumer protection education should also teach consumers how to check AI answers, keep records of consultations, request human intervention, and practice submitting re-evaluation requests themselves.

At the ISO Committee on Consumer Policy General Meeting in May, the discussion around digital consumer rights focused on comprehensibility, fairness, and access to remedies. These same three criteria must be used to check whether consumers in financial settings have actually understood the explanations given to them, were not unfairly excluded based on their circumstances, and were able to solve their issues through objection and re-evaluation processes.

The use of AI is expanding—from fraud detection to product review, sales, and complaint handling. The AI-driven financial consumer protection competence required of both the Financial Supervisory Service and financial firms is the ability to identify and resolve problems left unaddressed by algorithms, and to continuously improve work procedures and practices so that such inconveniences and harm are not repeated.

Jung Unyoung, Chairperson of the Finance and Happiness Network and Visiting Professor, Department of Sustainable Economics, Hanyang University

This content was produced with the assistance of AI translation services.