Abstract
This study investigates the critical dimensions of data quality influencing the quality of AI-supported decision-making processes in professional environments. Using a quantitative, cross-sectional research design, primary data were collected from 272 professionals via the Prolific platform. The conceptual model examined four data quality dimensions—Accuracy, Objectivity, Believability, and Reputation—and the moderating role of age, using Partial Least Squares Structural Equation Modeling (PLS-SEM) in SmartPLS 4.
The results indicate that Accuracy ( β= 0.387, p < 0.001) and Objectivity ( β= 0.178, p = 0.012) are significant predictors of decision-making quality, collectively explaining approximately 37% of the variance (R² = 0.369). Accuracy demonstrated a medium effect size, while objectivity showed a small effect. Conversely, the impacts of Believability and Reputation were not statistically significant. Furthermore, the study found no evidence that Age moderates these relationships, suggesting that the reliance on high-quality AI outputs is consistent across generational cohorts.
These findings contribute to Information Systems literature by identifying the paramount importance of technical correctness over source reputation in the context of Large Language Models (LLMs). For organizations, the study implies that AI implementation strategies should prioritize output precision and unbiased data, as these factors are the primary drivers of effective professional decision-making regardless of the user's age.