A psychometric analysis of the artificial intelligence skills scale developed through chat gpt

This study aims to examine the psychometric properties of a scale developed to measure teachers’ ability to use artificial intelligence tools in education. The scale was originally proposed as having 33 items across 5 sub-dimensions by Chat GPT-4, an AI system. To establish content validity, the scale form was submitted to expert review. An exploratory factor analysis (EFA) was then conducted on a sample of 347 teachers to determine if the 33-item, 5-subdimensional structure proposed by Chat GPT was confirmed by empirical data. The EFA results supported Chat GPT-4’s proposed structure. A confirmatory factor analysis (CFA) was applied to a new sample of 603 teachers to assess the scale’s validity in an independent sample. The CFA results did not fully support the structure consisting of 33 items and 5 subdimensions proposed by Chat GPT. Reliability was analyzed using Cronbach’s Alpha and test-retest methods.