Can We Assess Attitudes Toward AI with Single Items? Associations with Existing Attitudes Toward AI Measures and Trust in ChatGPT

Abstract

Abstract A growing number of researchers investigate individual differences in attitudes toward Artificial Intelligence (AI), which is not surprising given that the AI revolution is impacting societies around the globe. Different frameworks have been proposed to study both positive and negative attitudes toward AI. To our knowledge, the present work is the first to simultaneously investigate the ATAI (Attitudes Toward Artificial Intelligence Scale) and the GAAIS (General Attitudes Towards Artificial Intelligence Scale). Further, two single items assessing positive and negative attitudes toward AI were added to the study to see if they would grasp substantial parts of the variance of the already established ATAI and GAAIS inventories. Correlations were of moderate to large effect size when comparing associations between the single-item measures and both ATAI and GAAI scales ( German speaking sample 1 = 151 participants; German speaking sample 2 = 386). Finally, also associations with trusting the generative AI ChatGPT were included as external validation measurement in both investigated samples. Results revealed that all attitudes toward AI measures were associated with trusting ChatGPT. Moreover, a stepwise regression model demonstrated that the acceptance scale of the ATAI was the best predictor for trust in ChatGPT in sample 1, with more predictors in sample 2. The present work shows substantial overlap between the available attitudes towards AI measures, and this could be replicated in two samples. These insights can help future researchers and AI designers to choose the appropriate survey tool when considering to assess attitudes toward AI.

Ähnliche Arbeiten