Seung-Hee Kim, Journal of Internet Computing and Services, Vol. 25, No. 5, pp. 67-86, Oct. 2024
10.7472/jksii.2024.25.5.67, Full Text:  HTML
Keywords: artificial intelligence quality, ISO/IEC 25000, product quality, quality-in-use, Software Quality

Abstract

This study developed a novel ISO/IEC 25010 quality model for the quality management of artificial intelligence (AI)-based software by using quality characteristics classification card (QCCC) quality models. We used AI models to add, modify, and restructure AI quality attributes for the product quality model and the quality-in-use model of the ISO/IEC 25010 quality model to derive a novel ISO/IEC 25010 quality model. By integrating quality standards derived from various AI-related models, we enhanced the accuracy of the derived model. The product quality model included 10 main quality and 45 subquality attributes, and the quality-in-use model included 10 main quality and 28 subquality attributes. In AI-based models, the quality-in-use model was found to require modifications. The results revealed the direction of improvement of the AI-compatible software quality model and the possibilities for potential standardization and conflict resolution. This study presents the direction for standardization reviews on reorganizing the quality attributes, concepts of attributes, and relationships so that they can be applied to AI software while maintaining the framework of the currently defined software quality model. The results can serve as criteria for the quality management of AI-based software and can also contribute to research on quality models for AI-based software.

Statistics
Cite this article