Academic AI overreliance scale: development and preliminary psychometric validation in health sciences university...

The rapid integration of generative artificial intelligence (AI) into higher education has raised growing concerns regarding its potential impact on academic autonomy, self-regulated learning, and cognitive functioning. However, limited attention has been given to the progressive delegation of academic cognitive and self-regulatory processes to intelligent systems. This study aimed to develop and examine the initial psychometric properties of the Academic AI Overreliance Scale among health sciences university students.

Two independent studies were conducted in Ecuador. In Study 1, an Exploratory Factor Analysis (EFA) was performed with 437 university students to identify the underlying structure of an initial 24-item pool. In Study 2, a Confirmatory Factor Analysis (CFA) was conducted with an independent sample of 437 university students to evaluate the factorial validity, measurement invariance, and internal consistency of the refined 22-item version.

Study 1 supported a two-factor solution composed of Emotional-Dysregulated Reliance and Cognitive Reliance. Study 2 confirmed the correlated two-factor model, showing excellent fit indices (CFI = 0.994, TLI = 0.994, RMSEA = 0.066, SRMR = 0.056) and strict invariance across sex groups. Internal consistency was adequate for both Emotional-Dysregulated Reliance (α = 0.94, ω = 0.94) and Cognitive Reliance (α = 0.88, ω = 0.88).

The findings suggest that Academic AI Overreliance may represent a specific form of cognitive and self-regulatory externalization associated with the integration of generative AI into academic activities, rather than a traditional form of technological addiction. The scale provides preliminary evidence of validity based on internal structure, measurement invariance, and reliability, and may serve as an initial instrument for future research on generative AI, academic reasoning, and intellectual autonomy in higher education.