0
/ This image was generated by AI.

As establishing AI governance emerges as a top priority in the financial sector, recommendations have emerged that each financial company should be allowed to build its own suitable governance framework. The rationale is that since excessive regulation could stifle productivity and innovation in the financial sector, financial authorities should clearly set out minimum principles and standards needed for AI use, while leaving the specifics of management systems to each financial company's discretion.

Seo Jung-ho, a senior research fellow at the Korea Institute of Finance, stressed in a report titled "Policy Tasks for Building AI Governance at Financial Companies" that financial authorities need to concretely present core principles related to AI development and use, such as fairness, transparency, and consumer protection. He explained that ambiguous regulation can become an obstacle to both effective oversight and innovation, making it necessary to clearly define the minimum requirements for AI use.

However, he noted that regulation should be oriented toward performance-based supervision. While risks that can arise from AI use need to be managed, excessive preemptive regulation that stifles productivity and innovation in the financial sector should be avoided. He pointed out in particular that relying solely on after-the-fact measures — identifying and sanctioning those responsible once an AI-related financial incident occurs — makes it difficult to build an effective AI management system.

He stressed that the specific method of building governance should be left to each financial company's discretion. Since the nature and use of AI models varies by company, rather than applying uniform standards, each firm should be allowed to establish an appropriate management system. The key, he said, is clearly defining who is accountable in AI-related decision-making processes and establishing a system that can manage risk across the entire lifecycle, from model development through operation, monitoring, and retirement.

Indeed, major financial groups are moving quickly to build their own governance frameworks, including forming dedicated organizations and consultative bodies. KB Financial Group established an "AI Ethics Committee" at its core subsidiary KB Kookmin Bank, serving as the top decision-making body for the ethical use of AI and related decision-making. In June, it launched a "Group Cyber Security Center" and specified 31 risk categories, among other steps to strengthen its control framework. Shinhan Financial Group has similarly built AI governance, establishing an AI risk management system tailored to the business characteristics of each affiliate. It has set up six group-wide AI principles as its top-level governance standards, covering accuracy and stability, transparency and explainability, and value for customers and society.

Hana Financial Group launched the "Hana AI Leaders Forum," where executives from each affiliate meet quarterly to discuss improving efficiency in the group's AI development and application management. It also plans to launch group-wide AI governance and a top decision-making body. Woori Financial Group plans to build a governance framework and risk assessment process for AI governance through an external expert consulting engagement, starting in April this year. NH Nonghyup Financial Group also plans to establish group-wide AI governance by the end of this year.

Seo said policy authorities need to be mindful that excessive AI regulation doesn't end up undermining productivity gains and innovation in the financial sector, adding that since guidelines are, after all, a form of soft regulation premised on voluntary compliance, how they are adopted should also be left up to each financial company — and that sanctions should never be imposed simply on the grounds that a company failed to follow them.

                                                                                                           Chae Jong-il