Ambivalence in Schoolchildren’s Perceptions of Artificial Intelligence Technologies AI

This article examines the ambivalent attitudes of schoolchildren aged 14–17 towards generative artificial intelligence in a context where institutional school frameworks fail to keep pace with students’ everyday practices. The empirical base comprises two online focus groups (n = 19; participants from St. Petersburg, Novosibirsk, and Vladivostok) and the results of web scraping and cluster analysis (TF-IDF, K-means) of regional supplementary education platforms. The theoretical framework draws on Anthony Giddens’ structuration theory, complemented by the concepts of child agency and ethical symmetry from the sociology of childhood. The findings reveal that schoolchildren construct an “ambivalent norm of conditional acceptance”: they acknowledge AI as a useful tool for structuring knowledge and reducing anxiety yet constrain its use in order to preserve cognitive autonomy and academic integrity. The study identifies spontaneously emerging informal rules for distinguishing between “help” and “cheating,” as well as strategies for concealing the use of neural networks from teachers. An analysis of supplementary education programers reveals the dominance of technical tracks over ethically reflective components. Overcoming the spontaneity of schoolchildren’s AI use and the broader sociotechnical gap in education is seen to require the creation of spaces of ethical symmetry between adults and children, with a focus on developing AI meta-literacy.