Keyword: «checklist»

MEO is a challenge. The practice of organizing work with parents in a preschool educational organization using the opportunities of the digital educational environment.

Relevance of the study is motivated by the active integration of artificial intelligence (AI) in education and the simultaneous criticism from educators regarding the quality of AI-generated products. Lesson plan generation is highly demanded by teachers for time efficiency and instructional variety. However, the lack of effective tools for the substantiated evaluation of AI materials reduces their didactic value. In this context, the problem of developing evaluation criteria and analyzing the sensitivity of neural networks to prompts becomes highly relevant. The aim of the study is to assess the sensitivity of various AI platforms to prompt phrasing during the generation of foreign language lesson stages aimed at developing grammatical skills in lower secondary school (grades 7–8), and to determine the extent to which the generated outputs comply with established methodological standards. A specialized checklist was developed as a methodological framework and evaluation tool to capture key conditions for organizing grammatical material. The empirical data comprised outputs generated by five AI platforms (DeepSeek, Perplexity, Gigachat, ChatGPT, and Qwen). A comparative analysis was conducted using four prompts, each varying by specific methodological constraints, which allowed for the tracking of algorithmic adaptability. The results of the study revealed that all platforms exhibited varying degrees of sensitivity to prompt detail. It was found that none of the generated lesson plans fully complied with the established methodological requirements, and all required mandatory pedagogical revision. The Qwen platform demonstrated the greatest initial suitability for the task at hand, although its results also required refinement. Typical methodological shortcomings inherent in AI-generated content were identified. The theoretical significance lies in systematizing approaches to the evaluation of educational content generated by generative models and expanding the understanding of AI limitations in grammar instruction. The practical significance is the development of a checklist for educators to verify AI-generated materials. The study demonstrates that collaborative efforts between developers and methodologists are essential to improve neural network accuracy. Furthermore, it substantiates the approach of utilizing "imperfect" AI-generated lesson plans as materials for methodological analysis during pre-service teacher training and in-service professional development, thereby fostering teachers' critical thinking skills.