[ Download ]
Abstract
Generative artificial intelligence (GenAI)
may support audit work, but its association with audit quality depends on how
auditors use and evaluate its outputs. This study examines task–technology fit,
quality-oriented utilization, professional judgment quality, and digital
competence among 214 external auditors serving Jordanian listed industrial
companies. Respondent-level averaged indicators were analyzed using a
reflective measurement model and regression equations corresponding to PROCESS
Models 1, 4, and 7. Audit-firm size and professional certification were
included as covariates. Reflective discriminant validity was supported by HTMT
values of 0.587–0.660. Across 5,000 bootstrap samples, the indirect association
through professional judgment was 0.115 (95% CI [0.078, 0.159]). Conditional
indirect associations increased from 0.048 at low digital competence to 0.154
at high competence; the index of moderated mediation was 0.082 (95% CI [0.043,
0.126]). The findings are consistent with a professional-judgment pathway and a
competence-related boundary condition. They support further evaluation of task
suitability, verification practices, and competence development.
Cross-sectional self-reports, aggregated indicators, and incomplete independent
validation of formative composites limit causal interpretation and
generalization.
JEL classification numbers: M42, M15, O33.
Keywords:
Generative artificial intelligence, External audit quality, Task-technology
fit, Professional judgment, Digital competence, Jordan.