Anxiety in the Age of AI: Constructing a Tool to Assess Public Perceptions

Abstract

The rapid proliferation of artificial intelligence (AI) technologies has sparked widespread debate about their societal implications, prompting a need to understand public attitudes, particularly anxiety, toward AI. This study aimed to develop and validate a scale to measure AI-related anxiety, exploring its correlates across demographic groups. Using an online survey of 708 Romanian adults conducted in October 2024, we constructed a 14-item scale with a 5-point Likert format, balancing positive and negative statements. Psychometric validation, including structural equation modelling to address method effects, confirmed the scale’s reliability (Cronbach’s alpha > 0,7) and fit (CFI = 0,963, RMSEA = 0,053). Results revealed that AI anxiety varies by education, occupation, income, and region, with lower anxiety among the highly educated, self-employed, and high-income individuals. Convergent validity was supported by correlations with technological readiness (-0,411) and a summated AI anxiety score (0,698). These findings align with prior research on AI perceptions and underscore the role of information access and experience in shaping attitudes. Despite this scale offers a robust tool for assessing AI anxiety, with implications for tailoring AI adoption strategies to diverse populations. Future research should pursue longitudinal and qualitative approaches to deepen understanding of this evolving phenomenon.