Min-Kyu Kim
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Min-Kyu Kim has a total of 13 co-authors. Here are the top co-authors by number of shared publications:
- Ji-Bum Chung
Department of Civil, Urban, Earth, and Environmental Engineering, Ulsan National Institute of Science and Technology, Ulsan, Republic of Korea
7 shared publications - Tae-An Yoo
Graduate School of Carbon Neutrality
4 shared publications - Eunseon Park2 shared publications
- Seunghoo Jeong
Railroad Structure Research Department, Korea Railroad Research Institute, Uiwang-si, Republic of Korea
2 shared publications Han-Sol Jang
1 shared publicationHo-Suk Yang
1 shared publication
Latest Publications
- Open Access
- 7 July 2026
- Published by Elsevier in SSRN Electronic Journal
Regulatory Gaps in Environmental Impact Assessment of Hyperscale AI Infrastructure: A Comparative Study across 28 Jurisdictions
Hyperscale data centers have become critical infrastructure for global artificial intelligence expansion, yet regulatory systems designed to anticipate their environmental consequences remain fragmented and underdeveloped. This study evaluates how existing EIA frameworks across 28 jurisdictions govern hyperscale data center development. The analysis combines comparative mapping of statutory EIA thresholds, assessment of public accessibility of EIA documentation, development of eight robustness criteria, and multi-coder content analysis of 15 retrievable EIA reports. The criteria cover power usage effectiveness, water use, groundwater protection, greenhouse gas mitigation, electronic waste management, life cycle assessment, cumulative impacts, and community participation. Three findings emerge. First, only 15 of 28 jurisdictions produced publicly accessible EIA documentation, revealing substantial transparency deficits across existing regulatory systems. Second, regulatory robustness varied substantially across jurisdictions: Finland, France, the United States, Germany, Italy, and Canada performed most strongly, whereas Japan, South Korea, India, and Chile clustered in the lowest tier; electronic waste management, cumulative impacts, and life cycle assessment were the weakest dimensions overall. Third, several hyperscale AI infrastructure projects proceeded without publicly accessible environmental review, including in jurisdictions with active EIA legislation, exposing uneven regulatory oversight. Among the 28 jurisdictions, only Italy maintains a dedicated EIA screening framework specifically calibrated to hyperscale data center development. The findings indicate that existing EIA systems remain poorly adapted to the environmental profile of AI infrastructure. The study proposes a sector-specific anticipatory governance framework to strengthen transparency, cumulative impact assessment, and environmental accountability in hyperscale data center development.
- 7 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessRegulatory Gaps in Environmental Impact Assessment of Hyperscale AI Infrastructure: A Comparative Study across 28 Jurisdictions
Hyperscale data centers have become critical infrastructure for global artificial intelligence expansion, yet regulatory systems designed to anticipate their environmental consequences remain fragmented and underdeveloped. This study evaluates how existing EIA frameworks across 28 jurisdictions govern hyperscale data center development. The analysis combines comparative mapping of statutory EIA thresholds, assessment of public accessibility of EIA documentation, development of eight robustness criteria, and multi-coder content analysis of 15 retrievable EIA reports. The criteria cover power usage effectiveness, water use, groundwater protection, greenhouse gas mitigation, electronic waste management, life cycle assessment, cumulative impacts, and community participation. Three findings emerge. First, only 15 of 28 jurisdictions produced publicly accessible EIA documentation, revealing substantial transparency deficits across existing regulatory systems. Second, regulatory robustness varied substantially across jurisdictions: Finland, France, the United States, Germany, Italy, and Canada performed most strongly, whereas Japan, South Korea, India, and Chile clustered in the lowest tier; electronic waste management, cumulative impacts, and life cycle assessment were the weakest dimensions overall. Third, several hyperscale AI infrastructure projects proceeded without publicly accessible environmental review, including in jurisdictions with active EIA legislation, exposing uneven regulatory oversight. Among the 28 jurisdictions, only Italy maintains a dedicated EIA screening framework specifically calibrated to hyperscale data center development. The findings indicate that existing EIA systems remain poorly adapted to the environmental profile of AI infrastructure. The study proposes a sector-specific anticipatory governance framework to strengthen transparency, cumulative impact assessment, and environmental accountability in hyperscale data center development.
- Open Access
- 7 July 2026
- Published by Elsevier in SSRN Electronic Journal
Regulatory Gaps in Environmental Impact Assessment of Hyperscale AI Infrastructure: A Comparative Study across 28 Jurisdictions
Hyperscale data centers have become critical infrastructure for global artificial intelligence expansion, yet regulatory systems designed to anticipate their environmental consequences remain fragmented and underdeveloped. This study evaluates how existing EIA frameworks across 28 jurisdictions govern hyperscale data center development. The analysis combines comparative mapping of statutory EIA thresholds, assessment of public accessibility of EIA documentation, development of eight robustness criteria, and multi-coder content analysis of 15 retrievable EIA reports. The criteria cover power usage effectiveness, water use, groundwater protection, greenhouse gas mitigation, electronic waste management, life cycle assessment, cumulative impacts, and community participation. Three findings emerge. First, only 15 of 28 jurisdictions produced publicly accessible EIA documentation, revealing substantial transparency deficits across existing regulatory systems. Second, regulatory robustness varied substantially across jurisdictions: Finland, France, the United States, Germany, Italy, and Canada performed most strongly, whereas Japan, South Korea, India, and Chile clustered in the lowest tier; electronic waste management, cumulative impacts, and life cycle assessment were the weakest dimensions overall. Third, several hyperscale AI infrastructure projects proceeded without publicly accessible environmental review, including in jurisdictions with active EIA legislation, exposing uneven regulatory oversight. Among the 28 jurisdictions, only Italy maintains a dedicated EIA screening framework specifically calibrated to hyperscale data center development. The findings indicate that existing EIA systems remain poorly adapted to the environmental profile of AI infrastructure. The study proposes a sector-specific anticipatory governance framework to strengthen transparency, cumulative impact assessment, and environmental accountability in hyperscale data center development.
- 7 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessRegulatory Gaps in Environmental Impact Assessment of Hyperscale AI Infrastructure: A Comparative Study across 28 Jurisdictions
Hyperscale data centers have become critical infrastructure for global artificial intelligence expansion, yet regulatory systems designed to anticipate their environmental consequences remain fragmented and underdeveloped. This study evaluates how existing EIA frameworks across 28 jurisdictions govern hyperscale data center development. The analysis combines comparative mapping of statutory EIA thresholds, assessment of public accessibility of EIA documentation, development of eight robustness criteria, and multi-coder content analysis of 15 retrievable EIA reports. The criteria cover power usage effectiveness, water use, groundwater protection, greenhouse gas mitigation, electronic waste management, life cycle assessment, cumulative impacts, and community participation. Three findings emerge. First, only 15 of 28 jurisdictions produced publicly accessible EIA documentation, revealing substantial transparency deficits across existing regulatory systems. Second, regulatory robustness varied substantially across jurisdictions: Finland, France, the United States, Germany, Italy, and Canada performed most strongly, whereas Japan, South Korea, India, and Chile clustered in the lowest tier; electronic waste management, cumulative impacts, and life cycle assessment were the weakest dimensions overall. Third, several hyperscale AI infrastructure projects proceeded without publicly accessible environmental review, including in jurisdictions with active EIA legislation, exposing uneven regulatory oversight. Among the 28 jurisdictions, only Italy maintains a dedicated EIA screening framework specifically calibrated to hyperscale data center development. The findings indicate that existing EIA systems remain poorly adapted to the environmental profile of AI infrastructure. The study proposes a sector-specific anticipatory governance framework to strengthen transparency, cumulative impact assessment, and environmental accountability in hyperscale data center development.
- 12 June 2026
- Published by SAGE Publications in WORK: A Journal of Prevention, Assessment & Rehabilitation
Return-to-Work challenges faced by non-regular workers: Insights from a four-year panel study in South Korea
Background: Non-regular workers in South Korea account for approximately 37% of the workforce and face elevated risks of occupational injury, yet empirical research on their return-to-work (RTW) outcomes remains limited. Objective: This study aims to (1) compare RTW outcomes between regular and non-regular workers over a four-year observation period, and (2) identify group-specific predictors of RTW using a multidimensional framework encompassing individual, psychological, health-related, social, and workplace factors. Methods: Data from 2643 occupationally injured workers were drawn from the Panel Study of Workers’ Compensation Insurance, a nationally representative dataset spanning four annual waves (2018 to 2021). Random-effects panel multinomial logistic regression models were estimated separately for each employment type, supplemented by a pooled interaction model to test between-group differences. Results: Non-regular workers were more likely to remain unemployed or be reemployed at a different workplace, whereas regular workers were more likely to return to their original jobs. Self-efficacy and social interaction were associated with RTW in both groups, though their effects operated through different pathways. Labor union presence was significantly associated with RTW only among regular workers. Conclusions: Non-regular workers experience structurally distinct RTW patterns following occupational injury. The findings highlight the need for targeted interventions that strengthen psychological and social resources for non-regular workers, alongside institutional reforms to support their sustainable reintegration.
- 12 June 2026
- Published by SAGE Publications in WORK: A Journal of Prevention, Assessment & Rehabilitation
Return-to-Work challenges faced by non-regular workers: Insights from a four-year panel study in South Korea
Background: Non-regular workers in South Korea account for approximately 37% of the workforce and face elevated risks of occupational injury, yet empirical research on their return-to-work (RTW) outcomes remains limited. Objective: This study aims to (1) compare RTW outcomes between regular and non-regular workers over a four-year observation period, and (2) identify group-specific predictors of RTW using a multidimensional framework encompassing individual, psychological, health-related, social, and workplace factors. Methods: Data from 2643 occupationally injured workers were drawn from the Panel Study of Workers’ Compensation Insurance, a nationally representative dataset spanning four annual waves (2018 to 2021). Random-effects panel multinomial logistic regression models were estimated separately for each employment type, supplemented by a pooled interaction model to test between-group differences. Results: Non-regular workers were more likely to remain unemployed or be reemployed at a different workplace, whereas regular workers were more likely to return to their original jobs. Self-efficacy and social interaction were associated with RTW in both groups, though their effects operated through different pathways. Labor union presence was significantly associated with RTW only among regular workers. Conclusions: Non-regular workers experience structurally distinct RTW patterns following occupational injury. The findings highlight the need for targeted interventions that strengthen psychological and social resources for non-regular workers, alongside institutional reforms to support their sustainable reintegration.
- 21 March 2026
- Published by Taylor & Francis in Interactive Learning Environments
- Vol. 34 (8) , 1-17
The impact of AI-assisted writing on students’ learning effectiveness and perceptions: a mixed-methods study
- 21 March 2026
- Published by Taylor & Francis in Interactive Learning Environments
- Vol. 34 (8) , 1-17
The impact of AI-assisted writing on students’ learning effectiveness and perceptions: a mixed-methods study
- Top Cited PapersOpen Access
- 2 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 18881-18896
Toward Sustainable AI: A Scoping Review of Carbon Footprint and Environmental Impacts Across Training and Inference Stages
AI is spreading rapidly, creating significant social and economic value while also raising concerns about its high energy use and environmental sustainability. While prior studies have predominantly focused on the energy-intensive nature of the training phase, the cumulative environmental footprint generated during large-scale service operations, particularly in the inference phase, has received comparatively less attention and remains difficult to compare across studies due to inconsistent boundaries and disclosure practices. To bridge this gap, this study conducts a scoping review of methodologies and evidence on AI carbon footprint assessment. We analyze the classification and standardization status of existing AI carbon measurement tools and methodologies, and comparatively examine the environmental impacts arising from both training and inference stages, and synthesize inference evaluation into a benchmark-oriented framework. In addition, we identify how multidimensional factors such as model size, prompt complexity, serving environments, and system boundary definitions shape the resulting carbon footprint. Our review reveals critical limitations in current AI carbon accounting practices, including methodological inconsistencies, technology-specific biases, and insufficient attention to end-to-end system perspectives. Building on these insights, we propose future research and governance directions: (1) establishing standardized and transparent universal measurement protocols with clear boundary, functional-unit, and disclosure requirements, (2) designing dynamic evaluation frameworks that incorporate user behavior and service-level operating conditions, (3) developing life-cycle monitoring systems that encompass embodied emissions, and (4) advancing multidimensional sustainability assessment framework that balance model performance with environmental efficiency. This paper provides a foundation for interdisciplinary dialogue aimed at building a sustainable AI ecosystem and offers a baseline guideline for researchers and policymakers seeking to evaluate AI environmental impacts across technical, social, and operational dimensions.
- 2 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 18881-18896
Top Cited PapersOpen AccessToward Sustainable AI: A Scoping Review of Carbon Footprint and Environmental Impacts Across Training and Inference Stages
AI is spreading rapidly, creating significant social and economic value while also raising concerns about its high energy use and environmental sustainability. While prior studies have predominantly focused on the energy-intensive nature of the training phase, the cumulative environmental footprint generated during large-scale service operations, particularly in the inference phase, has received comparatively less attention and remains difficult to compare across studies due to inconsistent boundaries and disclosure practices. To bridge this gap, this study conducts a scoping review of methodologies and evidence on AI carbon footprint assessment. We analyze the classification and standardization status of existing AI carbon measurement tools and methodologies, and comparatively examine the environmental impacts arising from both training and inference stages, and synthesize inference evaluation into a benchmark-oriented framework. In addition, we identify how multidimensional factors such as model size, prompt complexity, serving environments, and system boundary definitions shape the resulting carbon footprint. Our review reveals critical limitations in current AI carbon accounting practices, including methodological inconsistencies, technology-specific biases, and insufficient attention to end-to-end system perspectives. Building on these insights, we propose future research and governance directions: (1) establishing standardized and transparent universal measurement protocols with clear boundary, functional-unit, and disclosure requirements, (2) designing dynamic evaluation frameworks that incorporate user behavior and service-level operating conditions, (3) developing life-cycle monitoring systems that encompass embodied emissions, and (4) advancing multidimensional sustainability assessment framework that balance model performance with environmental efficiency. This paper provides a foundation for interdisciplinary dialogue aimed at building a sustainable AI ecosystem and offers a baseline guideline for researchers and policymakers seeking to evaluate AI environmental impacts across technical, social, and operational dimensions.
- Open Access
- 4 July 2025
- Published by International Society of Global Health in Journal of Global Health
- Vol. 15, 04189
Political construction of risk perception and preventive behaviour during the COVID-19 pandemic in the Republic of Korea
Risks that are perceived as dreadful, uncontrollable, and unfamiliar tend to be regarded as dangerous, particularly when compared with other existing and anticipative risks [1,2]. Such perceptions of a specific risk can be amplified or diminished, influenced by factors such as hazard characteristics [1,2]. In addition to hazard characteristics, receiver traits can also significantly influence […]
- 4 July 2025
- Published by International Society of Global Health in Journal of Global Health
- Vol. 15, 04189
Open AccessPolitical construction of risk perception and preventive behaviour during the COVID-19 pandemic in the Republic of Korea
Risks that are perceived as dreadful, uncontrollable, and unfamiliar tend to be regarded as dangerous, particularly when compared with other existing and anticipative risks [1,2]. Such perceptions of a specific risk can be amplified or diminished, influenced by factors such as hazard characteristics [1,2]. In addition to hazard characteristics, receiver traits can also significantly influence […]
- Open Access
- 24 June 2025
- Published by Elsevier in Journal of Transport & Health
- Vol. 44
Risk perception, protective behavior, and transit ridership during COVID-19: Longitudinal insights from Seoul Metropolitan Area, Republic of Korea
- 24 June 2025
- Published by Elsevier in Journal of Transport & Health
- Vol. 44
Open AccessRisk perception, protective behavior, and transit ridership during COVID-19: Longitudinal insights from Seoul Metropolitan Area, Republic of Korea