Using Large Language Models to Create Educational Materials for the “Databases†Discipline

This paper examines the use of large language models (LLM) for creating educational materials. A practical methodology for generating high-quality educational content for the speciï¬ c discipline of “Databases†is proposed and veriï¬ ed. A multi-stage methodology is presented, in which one LLM generates content, and a second, independent “reasoning†model veriï¬ es its quality and correctness. A comparison method with an authoritative source and a modiï¬ ed “veriï¬ cation chain†algorithm was used to check the generated materials for factual errors. The results conï¬ rm that this approach, when used with modern, high-performance LLMs (such as DeepSeek and Gemini), enables the creation of high-quality educational texts with a low probability of hallucinations. The methodology can signiï¬ cantly accelerate the development of reliable educational materials and can be optimized by reducing the number of iterations while maintaining a high-quality initial response.