A Three-Layered Framework and Measurement Toolkit for Understanding Emotional and Relational Engagement with AI

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

As generative artificial intelligence (AI) becomes increasingly embedded in emerging adults’ academic, work, and social lives, interactions with AI increasingly involve emotional disclosure and relational engagement beyond instrumental use. However, empirical research has been constrained by the lack of validated instruments capturing behavioral, cognitive, and emotional aspects of human-AI engagement. This study developed and validated a three-scale measurement toolkit using a regionally representative sample of college students from 26 universities in Shanghai (N = 1406, Mage = 18.98, SD = 1.25, 59.7% female). The toolkit includes a Multidimensional AI Use Scale, a Perceived Uniqueness of AI Interaction Scale, and an AI Emotional Attachment Scale. Split-sample exploratory and confirmatory factor analyses supported clear factor structures, satisfactory reliability, and evidence of discriminant and criterion-related validity. Together, these instruments provide a psychometrically sound framework for examining behavioral use, cognitive perceptions, and affective attachment in human-AI engagement among emerging adults.