NPS CIO Seo Won-joo: "We will identify areas where AI can be used, including exchange-rate analysis" [fn Market Watch]

[Financial News] The National Pension Service (NPS) is integrating artificial intelligence (AI) directly into its fund-management operations. Its first project focuses on exchange rates. With overseas assets accounting for more than half of its portfolio, fluctuations in the won-dollar exchange rate have emerged as a key factor affecting fund returns. NPS plans to use AI to interpret macroeconomic indicators, news, research reports and other unstructured information, thereby improving the speed and precision of investment decisions.
According to investment banking industry sources on the 12th, the National Pension Service signed an MOU with LG AI Research in Magok, Seoul, the previous day to cooperate on AI technology for fund management. The partnership will combine NPS’s investment-management experience with LG AI Research’s model-development capabilities to jointly develop AI models and services specialized for financial operations, then validate and apply them in actual work.
The two organizations agreed to cooperate on developing and validating an AI-based exchange-rate analysis service specialized for fund management, jointly researching AI technologies related to investment and applying them to investment operations, and strengthening NPS’s AI capabilities through working groups and training programs. Rather than limiting the partnership to simple consulting, the two sides will work together through the entire process, from development to field validation.
The first joint project, an AI-based exchange-rate analysis service, will analyze the factors driving exchange-rate movements by combining macroeconomic indicators such as interest-rate differentials, the current-account balance and inflation with text data from news articles and research reports. The plan is to quantify vast amounts of information that are difficult for people to process and provide it to investment professionals, while gradually validating its practical usability.
The partnership is being viewed as a step forward for public-sector AI use, which had largely remained at the level of information searches and document summaries. Rather than simply adopting a general-purpose model, the organizations will jointly design a finance-specific model and validate its performance in pension-fund operations. Starting with exchange-rate analysis, NPS plans to identify additional areas for AI use, including risk management, research automation and greater efficiency in investment processes.
LG AI Research has introduced EXAONE-BI, a financial AI agent that provides prediction scores for changes in approximately 8,000 stocks listed in South Korea and the United States, along with expert-level explanations. It has continued to train its models on financial-domain data and enhance them through collaboration with Koscom Corporation and London Stock Exchange Group (LSEG). This effort also aligns with a broader trend among global pension funds, including APG Asset Management in the Netherlands and the Canada Pension Plan Investment Board (CPPIB), which have established separate data and AI organizations and integrated them into their investment operations.
Hwayoung Lee, an executive vice president at LG AI Research, said, "This is a meaningful opportunity for EXAONE’s capabilities in the financial sector to be validated in the investment operations of the country’s largest pension fund." Lee added, "Together, we will create a model case for leading the use of AI in the public sector."
Seo Won-joo, chief investment officer (CIO) of the National Pension Service, said, "Through this agreement, we will apply AI technology to actual operations and systematically validate its potential uses." He added, "Starting with exchange-rate analysis, we will continue to identify areas where AI can be used, create an environment that allows investment professionals to focus on investment decisions, and further enhance the expertise and efficiency of fund management."
[email protected] Kang Gu-gwi Reporter