Risk identification, quantification and governance strategies for generative artificial intelligence service
Lists
Only you see this
Abstract
Purpose The purpose of this paper is to identify the diverse range of risks associated with generative artificial intelligence (AI) at various stages, achieve effective quantification of these risks and develop targeted risk governance strategies tailored to the specific risk profiles of generative AI at each stage. Design/methodology/approach This study aims to use the work breakdown structure-risk breakdown structure methodology to systematically identify risks associated with generative AI. By considering both the frequency of risk occurrence and the severity of potential losses, the study uses triangular fuzzy numbers to quantify potential risks at various stages. Finally, by comparing the quantitative risk assessment outcomes, this paper explores comprehensive risk governance strategies for generative AI. Findings A staged risk identification system for generative AI has been established, enabling effective quantification of risks. Additionally, tailored risk governance strategies have been provided for different stages of generative AI development. Originality/value This paper refines and enhances the risk identification system for generative AI, elucidates key risk governance points at various stages of AI development and is crucial for ensuring the safety and reliability of generative AI throughout its entire service lifecycle.
BibTeX
@inproceedings{yang2025risk,
author = {Ming Yang and Jia Wang and Tao Zhang and Rong Jiang and Tilei Gao},
title = {{Risk identification, quantification and governance strategies for generative artificial intelligence service}},
booktitle = {Chinese},
year = {2025},
doi = {10.1108/cms-07-2024-0540},
}Loading related content...