Development and Validation of the Reliance on Artificial Intelligence in Medicine Scale (RAIM)

Abstract

This article presents the development and validation of the RAIM (Reliance on Artificial Intelligence in Medicine) scale, designed to assess individuals’ reliance on AI in healthcare. The construct was defined as a functional and epistemic reliance on AI systems, distinct from interpersonal trust, which presupposes moral agency and normative obligations. Based on expert-generated content, 60 initial items were evaluated for content validity, yielding 30 items retained using Lawshe’s CVR method. Four studies were conducted on statistically representative samples of Polish adults. Exploratory and confirmatory factor analyses confirmed a two-factor structure (the technological dimension and the social-holistic dimension of reliance on artificial intelligence in medicine). The final 15-item scale demonstrated excellent internal consistency (α = .963 / .833) and high test–retest reliability, both immediately (r = .956 / .925) and after two weeks (r = .912 / .832). Convergent validity with the AI Attitudes Scale (AIAS-4) was also confirmed (r = .714 / -.657). The RAIM is a reliable and valid instrument for assessing person’s reliance on AI in medical contexts.