GAAIS-J: Translation and Validation of the Japanese Version of the General Attitudes Toward Artificial Intelligence Scale

Abstract

The rapid integration of artificial intelligence (AI) into daily life highlights the need for culturally adapted tools to assess public attitudes. This study translated, culturally adapted, and validated the Japanese version of the General Attitudes toward Artificial Intelligence Scale (GAAIS-J) to enable cross-cultural research. The original GAAIS was translated, back-translated, and reviewed by experts for cultural and linguistic equivalence. A web-based survey of 3689 Japanese was conducted. Confirmatory factor analysis supported a two-factor model-positive and negative attitudes-after removing one low-loading item and allowing select item covariances, yielding acceptable fit indices. Reliability was high (ω = 0.94, α = 0.92), and average variance extracted values indicated satisfactory convergent validity. Construct validity was demonstrated through correlations with AI trust/distrust, generative AI usage frequency, socioeconomic status, gender, and Big Five traits. Positive attitudes were linked to higher trust, income, and openness; negative attitudes correlated with older age, lower education, and lower agreeableness. The GAAIS-J is a reliable, valid instrument for assessing AI attitudes in Japan and supporting international comparisons.

Ähnliche Arbeiten