Semerci Şahin Remziye, Özbay Özkan, Çınar Özbay Sevil, Durmuş Sarıkahya Selma
Child and Disease Nursing Department, Nursing Faculty, Koç University, Istanbul, Türkiye.
Distance Education Application and Research Center, Artvin Çoruh University, Artvin, Türkiye.
Eur J Pediatr. 2025 Aug 30;184(9):585. doi: 10.1007/s00431-025-06435-8.
The widespread adoption of generative artificial intelligence (AI) in education and daily life necessitates a deeper understanding of students' awareness and attitudes. However, there is a lack of appropriate and psychometrically validated tools to assess these constructs reliably. This study aimed to examine the validity and reliability of a newly developed scale measuring secondary school students' awareness of generative artificial intelligence in the Turkish context. This methodological study employed a cross-sectional, descriptive, and correlational design. The research was conducted with 444 secondary school students aged 14-18. Data were collected using a Demographic Information Form and the Generative Artificial Intelligence Awareness Scale. The study followed a three-stage process in line with the STROBE checklist: (1) item development, (2) expert evaluation for content validity, and (3) construct validity and reliability testing. Construct validity was assessed using Exploratory and Confirmatory Factor Analyses. Reliability was examined through Cronbach's alpha, corrected item-total correlations, and split-half reliability analysis. EFA revealed a three-factor structure: Basic Knowledge of Generative Artificial Intelligence, Positive Attitudes Toward Generative Artificial Intelligence, and Concerns About the Impacts of Generative Artificial Intelligence, explaining 54.29% of the total variance. CFA supported the model, yielding acceptable fit indices (χ/df = 3.792, RMSEA = 0.056, GFI = 0.905, CFI = 0.927, TLI = 0.919), indicating good model fit. The Cronbach's alpha for the total scale was 0.946, with subscale coefficients ranging from 0.822 to 0.925. Corrected item-total correlations ranged from 0.512 to 0.733. Split-half reliability for the first factor showed strong consistency (Spearman-Brown = 0.891, Guttman = 0.887, correlation between halves = 0.804). The findings demonstrate that the Turkish version of the generative artificial intelligenceawareness scale is a valid and reliable instrument for assessing students' knowledge, attitudes, and concerns about generative AI. What is Known: • Generative artificial intelligence (GAI) technologies are increasingly used by adolescents in educational settings, yet their awareness, ethical understanding, and usage patterns remain underexplored. • Existing measurement tools for AI literacy often target higher education students and lack specificity in assessing secondary school students' awareness of generative AI. What is New: • This study developed and psychometrically validated the first comprehensive scale to assess secondary school students' awareness of generative AI, including knowledge, attitudes, and ethical concerns. • The Generative Artificial Intelligence Awareness Scale (GAIAS) provides educators and policymakers with a reliable tool to evaluate and support adolescents' digital and ethical preparedness in the age of generative AI.
生成式人工智能(AI)在教育和日常生活中的广泛应用,使得有必要更深入地了解学生的认知和态度。然而,目前缺乏合适的、经过心理测量学验证的工具来可靠地评估这些构念。本研究旨在检验在土耳其背景下新开发的一个测量中学生对生成式人工智能认知的量表的效度和信度。这项方法学研究采用了横断面、描述性和相关性设计。研究对象为444名年龄在14至18岁之间的中学生。数据通过人口信息表和生成式人工智能认知量表收集。该研究按照STROBE清单遵循了三个阶段的过程:(1)项目开发,(2)内容效度的专家评估,以及(3)构念效度和信度测试。使用探索性和验证性因素分析来评估构念效度。通过克朗巴哈系数、校正后的项目总分相关性和分半信度分析来检验信度。探索性因素分析揭示了一个三因素结构:生成式人工智能的基础知识、对生成式人工智能的积极态度以及对生成式人工智能影响的担忧,解释了总方差的54.29%。验证性因素分析支持该模型,得到了可接受的拟合指数(χ/df = 3.792,RMSEA = 0.056,GFI = 0.905,CFI = 0.927,TLI = 0.919),表明模型拟合良好。总量表的克朗巴哈系数为0.946,子量表系数在0.822至0.925之间。校正后的项目总分相关性在0.512至0.733之间。第一个因素的分半信度显示出很强的一致性(斯皮尔曼 - 布朗系数 = 0.891,古特曼系数 = 0.887,两半之间的相关性 = 0.804)。研究结果表明,生成式人工智能认知量表的土耳其语版本是评估学生对生成式人工智能的知识、态度和担忧的有效且可靠的工具。已知信息:• 青少年在教育环境中越来越多地使用生成式人工智能(GAI)技术,但其认知、伦理理解和使用模式仍未得到充分探索。• 现有的人工智能素养测量工具通常针对高等教育学生,在评估中学生对生成式人工智能的认知方面缺乏特异性。新内容:• 本研究开发并通过心理测量学验证了首个全面的量表,以评估中学生对生成式人工智能的认知,包括知识、态度和伦理担忧。• 生成式人工智能认知量表(GAIAS)为教育工作者和政策制定者提供了一个可靠的工具,用于评估和支持青少年在生成式人工智能时代的数字和伦理准备。