This paper presents a comparative analysis between authentic English as a Second Language (ESL) learning materials and AI-generated texts produced by ChatGPT, a system based on large language models (LLMs). The study explores the potential of AI text generation as a tool for developing instructional materials at CEFR level A2.The research is based on two parallel corpora: a collection of 49 authentic texts from the Speak Out (2nd Edition) student book and a corresponding set of 49 texts generated by ChatGPT using controlled prompts aligned with the same topics and proficiency level. Using corpus analysis and text analysis methods, including n-gram analysis, part-of-speech tagging, and frequency distribution, the study examines differences in lexical diversity, syntactic structure, and overall text characteristics.The findings indicate that AI-generated texts tend to be longer and demonstrate greater lexical variety, with higher frequencies of adjectives, adverbs, and verbs. In contrast, authentic materials show more formulaic language and higher repetition of common structures. These results suggest that AI-generated content can serve as a flexible and scalable resource for ESL educators, supporting the creation of level-appropriate and customizable teaching materials.This work contributes to the growing field of artificial intelligence in education by providing empirical evidence on the linguistic properties of AI-generated texts and their potential application in language learning contexts.
Evgeni Stanchev (2024) studied this question.