探討人工智慧科技導入情境下審計效率與審計品質之影響因素

Exploring the influencing factors of audit efficiency and audit quality in the context of the introduction of artificial intelligence

本研究旨在探討人工智慧(Artificial Intelligence)科技導入情境下審計效率與審計品質之影響因素。主要係結合Davis, Bagozzi, and Warshaw(l989)科技接受度模型(Technology Acceptance Model)之知覺易用性及知覺有用性構面，與DeLone and McLean(1992; 2003)資訊系統成功模型(Information System Success Model)之使用滿意度等作為中介構面，並納入該模型之三項資訊系統品質作為外部影響構面進行探討。本研究除探討AI資訊系統品質是否透過審計人員對於AI科技接受度及使用滿意度，進而影響審計效率與審計品質外，並利用多群組分析(Multi-Group Analysis)探討不同AI類型對於使用滿意度及審計效率及審計品質之影響。本研究係運用間卷調查蒐集資料，利用偏最小平方法的結構方程模型(Partial Least Squares Structural Equation Modeling)進行分析。研究結果顯示，AI的三項品質構面對知覺易用性、知覺有用性及使用滿意度皆有顯著影響，且進一步影響審計效率與審計品質。在輔助型與增強型AI使用者間，AI品質對知覺有用性與知覺易用性之影響力存在顯著差異，其中增強型AI相較於輔助型AI，更能有效提升AI品質對知覺有用性與知覺易用性的影響力。

This study aims to explore the factors influencing audit efficiency and audit quality in the context of the introduction of Artificial Intelligence (AI) technology. It primarily combines the constructs of perceived ease of use and perceived usefulness from the Technology Acceptance Model (TAM) by Davis, Bagozzi, and Warshaw (1989), with user satisfaction from the Information System Success Model by DeLone and McLean (1992; 2003) as mediating constructs. Additionally, it incorporates the three quality dimensions of information systems from the model as external influencing factors. This research investigates whether the quality of AI information systems affects audit efficiency and audit quality through auditors' acceptance of AI and user satisfaction. Furthermore, it employs Multi-Group Analysis to examine the impact of different types of AI on user satisfaction, audit efficiency, and audit quality. Data is collected through a questionnaire survey and analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM). The results indicate that the three quality dimensions of AI significantly influence perceived ease of use, perceived usefulness, and user satisfaction, which in turn affect audit efficiency and audit quality. Among users of assistive and augmentative AI, there is a significant difference in the impact of AI quality on perceived usefulness and perceived ease of use, with augmentative AI more effectively enhancing the influence of AI quality on these perceptions compared to assistive AI.