Measuring the credibility of generative AI-produced information: An exploratory factor analysis

Abstract

Information credibility is a central concept in information science research and is closely linked to how people evaluate and use information. Credibility perceptions also play an important role in the adoption and continued use of technologies such as generative AI (gen AI). As gen AI becomes increasingly integrated into everyday life, it is essential to understand how people perceive the credibility of information produced by these systems. As part of a larger survey study on the use (and non-use) and perceptions of gen AI in higher education conducted in mid-2024 and mid-2025, we evaluated a set of questionnaire items to assess the perceived credibility of information created by gen AI. In this paper, we present the questionnaire items, descriptive statistics, and correlation matrices from the two surveys, and the results of our exploratory factor analysis examining the underlying structure of the credibility measure. Across both datasets, we identified two distinct factors - output credibility, which refers to the perceived credibility of AI-produced output itself, and relative credibility, which refers to perceptions of AI-produced output relative to human-produced information. We share the instrument and findings to support future refinement and adaptation in studying credibility in the context of gen AI.