Contemporary AI attitude scale research: a systematic mapping review and bibliometric analysis.

INTRODUCTION

The rapid expansion of artificial intelligence (AI) has been accompanied by growing efforts to measure individuals' attitudes toward these technologies. However, existing AI attitude scales vary considerably in their theoretical foundations, dimensional structures, target populations, and application contexts, raising questions about the extent to which they represent equivalent conceptualizations of the construct.

METHODS

This systematic mapping review examined contemporary AI attitude scale research and mapped its thematic, intellectual, methodological, and conceptual structure. A predefined Web of Science Core Collection search identified 298 eligible AI-related psychometric scale studies published between 2022 and June 2026; within this corpus, AI attitude was the most frequently represented construct, with 51 studies forming the final analytic sample. The review integrated descriptive and conceptual synthesis with bibliometric mapping analyses, including author keyword and abstract-term co-occurrence, co-citation, and bibliographic coupling.

RESULTS

The findings showed that scale development and psychometric evaluation constitute prominent methodological themes in the literature. Co-citation analysis linked the broader technology-acceptance tradition with AI-specific approaches to attitude measurement, while bibliographic coupling revealed four interconnected contemporary research fronts involving scale refinement and contextual extension, education and generative AI, population- and domain-specific measurement, and cross-cultural validation around established scale families. Despite these overlapping knowledge bases, substantial conceptual heterogeneity was evident across instruments, including unidimensional, positive-negative, cognitive-affective-behavioral, functional, ethical, and domain-specific multidimensional conceptualizations.

DISCUSSION: Overall, the findings indicate that bibliographic and methodological commonality has not produced conceptual uniformity in AI attitude measurement. Future research should therefore prioritize explicit construct definition, conceptual fit between instruments and research purposes, measurement invariance, cross-cultural validation, and cumulative evaluation of established measures rather than assuming equivalence across scales carrying the same construct label.