Development of the GAI Dependence Scale: a validity and reliability study

While individuals use generative artificial intelligence (GAI) to enhance their productivity, they may also develop psychological and behavioral over-reliance on these tools. Evaluating the extent of an individual’s dependence on GAI is crucial for conducting related research. The research involved 627 participants. Exploratory and confirmatory factor analysis was conducted to determine the construct validity of the scale. The results supported a structure consisting of four sub-dimensions and 19 items. The sub-dimensions are Usage Intensity, Content Dependence, Withdrawal Symptoms, and Negative Consequences. Confirmatory factor analysis indicated good model fit. Significant relationships were found with the GAI Dependence Scale, AI Dependence Scale (version of six items), and Behavioral Intention Scale during the analysis of the scale’s criterion validity. Cronbach’s α internal consistency, and the test-retest method were used to assess the reliability of the scale. Cronbach’s α internal consistency coefficient for the total score was found to be .92 for Sample 1 and .90 for Sample 2. Additionally, the test-retest reliability over a four-week interval yielded a coefficient of .81. Both coefficients are considered to indicate acceptable levels of reliability. The scale is an effective and reliable tool for assessing individual dependence on GAI.