Dataset-649

The target population consisted of adult university students with prior experience using AI for learning. For the main survey, 750 responses were collected through university-based recruitment and the Wenjuanxing online survey platform, of which 649 valid responses were retained after data-quality screening.</p><p dir="ltr">All constructs were measured using established scales adapted to the AI-assisted learning context and assessed on seven-point Likert scales.

The variables include AI use behavior, cognitive vigilance, cognitive offloading, perceived AI-assisted task performance, higher-order thinking skills, academic self-efficacy, and demographic characteristics. Data were analyzed using partial least squares structural equation modeling (PLS-SEM) in SmartPLS 4. Measurement reliability and validity were assessed before testing direct, mediating, and moderating effects using bootstrapping with 5,000 resamples.</p><p dir="ltr">This study was conducted in accordance with the Declaration of Helsinki.