Effect of Thinking Styles and Perception Modalities on Schoolchildren’s Ability to Identify AI-generated Content

Thinking and cognitive styles may affect the ability of schoolchildren to identify whether content was generated by humans or AI. This psychological study examined the integration of neural networks in education, aiming to detect the dominant thinking styles and perception modalities that help students identify AI-generated content. The experiment involved 122 schoolchildren aged 15–16 from a comprehensive secondary school in Novokuznetsk. The participants completed the following questionnaires: the Representational System Bias Test (B. Lewis, R. Pucelik), the Thinking Profile Questionnaire (V. A. Ganzen, K. B. Malyshev, L. V. Ogineis), and the Inquiring Mode Questionnaire (InQ) (R. Bramson, A. Harrison). Statistical analysis using the Kruskal-Wallis test revealed significant differences between groups. Subsequent post-hoc analysis identified specific groups that differed significantly in their ability to determine content origin. Digitals were significantly more successful at identifying content origin than kinesthetics, auditories, and visuals. Furthermore, students with a sign-based thinking type outperformed those with iconic, object-based, or symbolic thinking. Finally, individuals with an idealist thinking style were significantly more effective than synthesists, realists, and analysts. The research confirms the need to consider students’ cognitive characteristics when integrating AI-related technologies into the educational process.