Paper Title
A Validity and Reliability Study of AI Literacy Scales for Civil Servants

Abstract
This study aimed to develop and validate a scale for assessing civil servants' AI literacy. To address the unique characteristics of public sectors, such as non-profit performance evaluation and bureaucratic operational structures, this study integrated competency-based and psychological cognition component to tailor an AI literacy scale specifically for civil servants. Based on the literature, the initial version of the scale included eight constructs with 40 items: Understand AI, Apply AI, AI Ethics, Detect AI, Assistance, Anthropomorphic Interaction, Comfort with AI, and Technological Skills. After validation with 99 samples, the constructs Detect AI and Anthropomorphic Interaction were removed, resulting in a final scale with six constructs which are Understand AI, Apply AI, AI Ethics, Assistance, Comfort with AI, and Technological Skills and 20 items. The findings revealed that civil servants scored highest in Assistance and lowest in AI Ethics, with significant variations observed across age groups (39 to 49 and 50+ years). Additionally, the Master's degree group scored significantly higher than the Bachelor's degree group in the constructs of Apply AI and Understand AI. These results highlight the need for targeted AI literacy training to empower civil servants to adapt to dynamic environments. This research provides a practical tool to support the development of AI competencies in the public sector, thereby ensuring improved service outcomes for citizens. Keywords: Artificial Intelligence, AI Literacy, AI self-efficacy, civil servants training, public sector human resource development, Adult learning