Artificial Intelligence Myths Scale (AIM Scale): development and psychometric validation

Misconceptions about artificial intelligence (AI) are widespread in society, often leading to unfounded fears and resistance that hinder AI's acceptance and effective integration across various fields. These myths can negatively impact public understanding and the development of AI-related policies and education. Despite their significance, to the best of our knowledge, there is currently no psychometrically validated instrument specifically designed to systematically measure prevalent myths about AI in Türkiye and in the broader international literature. To address this gap, the present study aimed to develop and psychometrically validate the Artificial Intelligence Myths Scale (AIM Scale) within the Turkish cultural context, contributing to research on misconceptions and social perceptions related to AI.

The study was conducted in three phases. Study 1 involved exploratory factor analysis (EFA) on data from 453 undergraduate students to explore the scale's factor structure. Study 2 employed confirmatory factor analysis (CFA) with an independent sample of 453 participants to verify the identified model in Study 1. The scale's internal consistency and reliability were assessed using Cronbach's α, ordinal α, McDonald's ω coefficients, and corrected item-total correlations. In Study 3, Pearson correlation analysis was performed to examine whether the level of artificial intelligence myths and artificial intelligence literacy variables has a statistically significant correlation. Ethical approval was obtained, and informed consent was secured from all participants.

EFA results supported a unidimensional factor structure comprising 17 items. CFA confirmed this model with acceptable to good fit indices (χ²/df = 3.882, CFI = 0.991, TLI = 0.990, RMSEA = 0.080, 95% CI [0.072, 0.088], SRMR = 0.057). The scale demonstrated high reliability, with Cronbach's α = 0.934, ordinal α = 0.948, McDonald's ω = 0.939, and AVE = 0.531. Item-total correlations ranged from 0.548 to 0.774, indicating strong item discrimination.

The AIM Scale provides evidence supporting an instrument to assess AI-related myths within the Turkish undergraduate context, pending further cross-cultural validation. It provides a practical tool for researchers, educators, and policymakers to identify and address misconceptions about AI, thereby facilitating targeted educational interventions and improving public understanding. Further studies are needed to replicate these findings in more diverse demographic groups and to evaluate the scale's performance across different contexts.