AI literacy and critical evaluation of generative AI content: structure and gaps in future teachers’ competence

Abstract: Problem. Generative artificial intelligence has become an everyday tool for future teachers before it became a subject of professional training. Since its content is persuasive in form but opaque in source and depends on the wording of the prompt, spontaneous experience of working with it forms only individual components of AI literacy and critical AI evaluation skills; which ones remain deficient has not been established empirically. Aim. To identify the structure of AI literacy and critical AI evaluation skills of future teachers in evaluating generative AI content, which develops outside pur-poseful instruction; to determine their deficient components requiring special formation. Methods. A diagnostic survey of 141 students majoring in education at Pitirim Sorokin Syktyvkar State University was carried out in June 2026 via a chatbot on the VKontakte social network. The author’s questionnaire included 21 closed-type tasks, fifteen of which form a competence scale (maximum score – 15, KR-20 = 0.69), correlated with five components of AI literacy and critical AI evaluation skills. The groups were compared using the Mann–Whitney U-test with Holm correction. Results. The mean value of the final score was 8.36 (maximum – 15), which corresponds to 55.7 % of model answers. The reflective-critical component is developed noticeably more strongly than the others (76.8 %), while the value-motivational (40.9 %) and cognitive (42.6 %) components are the least developed. The score is higher among those who master techniques for con-structing structured prompts (9.88 versus 7.44, where p = 0.0004) and among those who verify the AI-generated result (8.80 versus 6.06, where p = 0.0018). Conclusions. Spontaneous experience develops wariness and the habit of verifica-tion, but it provides neither knowledge about the nature of generation errors, nor techniques for managing the prompt, nor an attitude towards open declaration of AI use. The value-motivational and cognitive components require priority devel-opment, building on the existing reflective-critical component.