Artificial intelligence (AI) is poised to transform the core functions of central banks, with experts suggesting that these institutions must undergo fundamental changes to fully leverage its potential. AI can be utilized in various areas, including monetary policy formulation and financial stability monitoring.
Sophia Kazinik, a senior researcher at Stanford University's Digital Economy Research Institute, presented these insights on June 2 during the 2026 BOK International Conference held at the Bank of Korea in Seoul.
Kazinik stated, "In monetary policy, AI can enhance the accuracy of real-time predictions by extracting and filtering high-frequency data, such as online product price listings and satellite image analysis, to complement the lag in official statistics. In terms of financial stability, large-scale unstructured text analysis can proactively detect signals of systemic risk."
Despite the potential benefits, the adoption of AI in global central banks has been slow due to the complexity of public objectives and the stringent accountability required. Sharing sensitive information externally poses additional challenges. In contrast, the Bank of Korea has already implemented its sovereign AI, 'BOKI,' on its internal network, becoming the first among global central banks to do so. BOKI is currently used for tasks such as report translation, regulation searches, and policy support.
During the presentation, experts emphasized the necessity of building integrated databases, expanding computing infrastructure, and securing access to computational resources to promote AI utilization. Kazinik noted, "Given the nature of central bank operations, a certain level of data barriers for privacy and governance control is unavoidable. However, to resolve bottlenecks in large-scale data usage and conduct advanced analytics, establishing a 'data lake' is essential."
When the U.S. Federal Reserve implemented AI in its operations, it was found to significantly enhance productivity. For instance, the New York Fed's open market operations (OMO) alone saved approximately 1.17 million hours annually. Treasury Services at the St. Louis and Atlanta Feds are projected to save around 3.18 million hours, while cash operations managed by the San Francisco, Kansas City, and Dallas Feds are expected to reduce labor hours by 3.51 million through AI.
Kazinik recommended that for successful AI integration, central banks should designate AI representatives within departments to foster an organic learning culture rather than solely relying on external experts. He emphasized, "For central banks to fully realize AI's potential, they must fundamentally innovate not only the relevant infrastructure but also their operational processes, systems, and norms alongside the technology."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.