BOK becomes first central bank to deploy in-house AI

Naver-built AI model enhances data security, operational efficiency for monetary authorities

Bank of Korea Gov. Rhee Chang-yong (left) and Naver founder and board chair Lee Hae-jin attend a joint conference held at the central bank’s headquarters in Seoul, Wednesday. (Im Se-jun/The Korea Herald)

The Bank of Korea and internet giant Naver on Wednesday unveiled an artificial intelligence model built exclusively for the central bank, marking what officials described as the first deployment of an in-house AI model by a central bank anywhere in the world.

The system, dubbed Bank of Korea Intelligence, or Boki, operates entirely on the central bank’s internal on-premise network, distinguishing it from cloud-based or externally hosted AI tools used by other monetary authorities.

“As one of the few countries capable of developing its own AI models, alongside the US and China, Korea places strong policy importance on fostering an ecosystem that supports the sustainable growth of its AI industry,” BOK Gov. Rhee Chang-yong said at a joint conference at the central bank’s headquarters in Seoul.

The homegrown platform represents a form of sovereign AI that strengthens Korea’s technological autonomy and control over sensitive financial data, the BOK explained.

“In this regard, the BOK’s collaboration with domestic tech firm Naver is expected to serve as a catalyst for strengthening the nation’s AI industry through public-private cooperation,” Rhee added.

Boki offers five core functions: research assistance, compliance support, document analysis, data analytics and translation of publications.

The BOK and Naver began developing Boki in 2024. Under the partnership, the tech company provided cloud infrastructure and large language models, while the central bank focused on developing AI applications tailored to financial and economic analysis.

While BOK staff previously had only limited access to external AI chatbots such as ChatGPT due to security concerns, the launch of Boki is expected to allow employees to use AI more proactively when handling sensitive data, as the model operates entirely within the central bank’s internal network.

More broadly, the BOK said the system could support policy analysis related to interest rate decisions by helping staff analyze large volumes of economic data and historical policy materials. Such inputs are central to deliberations by the Monetary Policy Board, which bases its decisions on a wide range of macroeconomic indicators.

Naver founder and board chair Lee Hae-jin said the project could enhance Korea’s competitiveness by strengthening its analytical capabilities in the economic and financial sectors.

“The BOK’s vast trove of data, which underpins the national economy, is itself a critical strategic asset for Korea,” Lee said. “We hope this AI model will go beyond transforming the BOK’s work culture and serve as a foundation for elevating the country’s financial and economic analysis capabilities and overall national competitiveness.”

The BOK’s partnership with Naver also enabled the central bank to lay the groundwork for digitizing its extensive document archive, allowing the AI model to learn from data assets accumulated through past policy deliberations and decision-making records.

According to Naver Cloud CEO Kim Yu-won, although only about 3 percent of the BOK’s documents had previously been digitized, the project paved the way for the central bank to standardize roughly 1.4 million internal documents into formats that can be understood and utilized by AI systems.

“By digitizing Korea’s uniquely handwritten documents containing Chinese characters, we have created a foundation for AI to study and analyze the central bank’s assets from past decisions,” Kim said.

The conference concluded with a demonstration of the AI system, presentations on security and governance enhancements, case studies on public- and financial-sector applications, and a panel discussion on the implications of sovereign AI and the broader expansion of AI-driven transformation.