Government Capability and Performance in the Era of Artificial Intelligence

• In Event: Artificial Intelligence and Smart Policymaking in South Korea

Abstract

Artificial intelligence (AI) has become a driving force for digital innovation across various sectors,

including government, where it promises to improve decision-making, deliver personalized services,

increase administrative efficiency, and enhance citizen engagement. Although nations worldwide,

including Korea, continue to expand public-sector AI investments, successful adoption does not

automatically yield improved government performance. Examples such as SyRI in the Netherlands

and COMPAS in Wisconsin highlight the need for a more comprehensive approach that encompasses

ethical and organizational factors in addition to technology.

This study’s primary objective is to develop a multidimensional understanding of government AI

capability, focusing especially on how it influences government performance and interacts with AI’s

technical features. While previous research often concentrated on legal, institutional, or acceptance

issues, few studies examined how organizations can effectively build AI-related capacity. Mikalef and

Gupta (2021) provided a resource-based framework detailing tangible, intangible, and human

resources. However, they did not fully capture ethical considerations vital to the public sector.

Addressing this gap, the current study integrates an ethical dimension, supported by Floridi et al

(2018) principles of beneficence, non maleficence, autonomy, justice, and explicability, into a

resource based model tailored for public organizations.

The study will also consider how AI-specific attributes like bias, transparency, instrumentality, and

effectiveness affect the transition from AI capability to performance outcomes. Under favorable

conditions, strong AI resources and ethics can yield significant benefits, yet certain technical

constraints or bias concerns may reduce these gains. By analyzing the interactive effects of

government AI capability and AI’s technical features, the study aims to produce a more nuanced

explanation of how performance outcomes materialize.

Methodologically, the research proceeds in two main stages. First, a comprehensive AI capability

model is developed by drawing on theory, expert feedback, and stakeholder input. A survey

instrument will be refined, then tested through factor analysis and reliability checks to confirm its

validity. Second, statistical analyses will determine the direct impact of AI capability on government

performance and examine the moderating role of AI’s technical traits.

Academically, this study seeks to expand our understanding of whether ethical conditions are

necessary in AI governance models. Furthermore, it aims to clarify how AI capability is linked to

government performance, and how AI-specific features influence this relationship. Practically, the

findings can guide policymakers in prioritizing AI investments, ensuring ethical safeguards are firmly

in place, and ultimately promoting public trust. By integrating ethical considerations into AI

implementation strategies, governments may mitigate potential risks and better achieve meaningful

results that serve the public interest.

Author

• Yejong Park, Yonsei University
  Presenting Author