@article{ART003283702},
author={Sung-Jin Kim and Kim Sun Jib},
title={Design of a Multimodal and Smart Contract-Based Real-Time Safety Evaluation System for High-Risk AI},
journal={Journal of Internet of Things and Convergence},
issn={2466-0078},
year={2025},
volume={11},
number={6},
pages={24}
TY - JOUR
AU - Sung-Jin Kim
AU - Kim Sun Jib
TI - Design of a Multimodal and Smart Contract-Based Real-Time Safety Evaluation System for High-Risk AI
JO - Journal of Internet of Things and Convergence
PY - 2025
VL - 11
IS - 6
PB - The Korea Internet of Things Society
SP - 24
EP -
SN - 2466-0078
AB - The recent proliferation of Artificial Intelligence(AI) in high-risk sectors, such as healthcare and finance, necessitates rigorous safety assessments to address data bias, security vulnerabilities, and ethical concerns. Although the National Information Society Agency(NIA) currently conducts safety certifications based on a 7-step process, the framework faces structural limitations, including static assessment methods, processing delays of one to two weeks caused by manual reviews, and a lack of transparency in audit logs. To address these issues, this study proposes a 9-step multimodal real-time safety assessment system that extends the existing NIA framework. Utilizing AWS Lambda, Amazon Comprehend, and Amazon Rekognition, the proposed system performs simultaneous analysis of text and images with a 5-minute bias detection cycle. This multimodal AI safety assessment approach achieved a 98% accuracy rate in information bias detection, representing an 8.89% improvement over existing research. Furthermore, by implementing automated logging for the entire process via Smart Contracts on Hyperledger Fabric, the system demonstrates six times faster response speeds and a 23% improvement in accuracy compared to conventional methods. This study is expected to contribute significantly to preventing data privacy breaches and enhancing safety assessments for high-risk AI applications.
KW - High-Risk AI;AI Safety Assessment;Multimodal Detection;Smart Contract;Real-time Bias;Detection;Blockchain
DO -
UR -
ER -
Sung-Jin Kim and Kim Sun Jib. (2025). Design of a Multimodal and Smart Contract-Based Real-Time Safety Evaluation System for High-Risk AI. Journal of Internet of Things and Convergence, 11(6), 24.
Sung-Jin Kim and Kim Sun Jib. 2025, "Design of a Multimodal and Smart Contract-Based Real-Time Safety Evaluation System for High-Risk AI", Journal of Internet of Things and Convergence, vol.11, no.6 24.
Sung-Jin Kim, Kim Sun Jib "Design of a Multimodal and Smart Contract-Based Real-Time Safety Evaluation System for High-Risk AI" Journal of Internet of Things and Convergence 11.6 24 (2025) : 24.
Sung-Jin Kim, Kim Sun Jib. Design of a Multimodal and Smart Contract-Based Real-Time Safety Evaluation System for High-Risk AI. 2025; 11(6), 24.
Sung-Jin Kim and Kim Sun Jib. "Design of a Multimodal and Smart Contract-Based Real-Time Safety Evaluation System for High-Risk AI" Journal of Internet of Things and Convergence 11, no.6 (2025) : 24.
Sung-Jin Kim; Kim Sun Jib. Design of a Multimodal and Smart Contract-Based Real-Time Safety Evaluation System for High-Risk AI. Journal of Internet of Things and Convergence, 11(6), 24.
Sung-Jin Kim; Kim Sun Jib. Design of a Multimodal and Smart Contract-Based Real-Time Safety Evaluation System for High-Risk AI. Journal of Internet of Things and Convergence. 2025; 11(6) 24.
Sung-Jin Kim, Kim Sun Jib. Design of a Multimodal and Smart Contract-Based Real-Time Safety Evaluation System for High-Risk AI. 2025; 11(6), 24.
Sung-Jin Kim and Kim Sun Jib. "Design of a Multimodal and Smart Contract-Based Real-Time Safety Evaluation System for High-Risk AI" Journal of Internet of Things and Convergence 11, no.6 (2025) : 24.
