Presentation (CTI)

OBJECTIVES: To examine the validity and reliability of two instruments measuring attitudes towards AI: the General Attitudes Towards Artificial Intelligence Scale (GAAIS) and the Attitudes Towards Artificial Intelligence (ATAI).
METHODS: Psychometric properties were assessed for reliability, validity, and responsiveness. Internal consistency was evaluated using Cronbach’s α, and test-retest reliability using intraclass correlation coefficients between baseline and 6-month follow-up among participants with stable AI attitudes. Factorial validity was examined using exploratory and confirmatory factor analyses. Concurrent validity was assessed using Pearson’s correlations between subscales. Responsiveness was evaluated using standardised mean differences among participants reporting changes in AI attitudes over time. This study further explored the attitude subscales and the factors associated with them, including sociodemographic characteristics and other variables of interest.
RESULTS: A survey of 1,076 adults in Australia, with 426 completing a retest between March and September 2023, confirmed the original factor structures of both scales. The ATAI captured Fear and Acceptance factors, while the GAAIS measured Negative and Positive factors. Model re-specifications improved fit, and test-retest reliability was strong overall, particularly for ATAI (4/5 items) and GAAIS (12/20 items). Intention to use AI or prior experience with AI consistently were significantly associated with more acceptance and positivity, and less fear and negativity. Gender, education, and age also associated with attitudes towards AI: males and those with a bachelor’s degree or higher reported greater acceptance and positivity, while older adults (60+) showed lower fear or negativity and higher positivity than younger adults.
CONCLUSIONS: These results support the ATAI and GAAIS as robust tools for assessing attitudes towards AI in Australia.