MEASURING AI LITERACY: A REVIEW OF GENERAL-PURPOSE SCALES AND A SIX-DIMENSIONAL FRAMEWORK

Abstract

The rapid development of generative artificial intelligence (AI) has made AI literacy one of the core competencies of citizens in the digital age. Yet current instruments for assessing AI literacy tend to be regionally bounded, lack cross-cultural equivalence, and measure only a narrow range of dimensions, which limits their usefulness for general assessment across cultures and contexts. This paper reviews Chinese- and English-language studies published between 2021 and 2025, combining CiteSpace bibliometric mapping with thematic analysis of 147 screened records, in order to compare the theoretical foundations, dimensional structures, psychometric properties, and cross-cultural adaptability of AI attitude scales (ALS) and generative AI literacy scales (GLAT). The review finds that most existing scales follow a four-dimensional model of cognition–skill–affect–ethics, but rarely address how assessment procedures might be standardized or how cross-cultural adaptation might work in practice. On this basis, the paper proposes a six-dimensional model—cognition, skill, affect, ethics, assessment, and cross-cultural—to provide theoretical support for building a general-purpose AI literacy scale, and it outlines directions and trends for future scale development.