APPLIED RESEARCH

Adaptation of authentic texts by means of generative artificial intelligence for teaching Russian as a foreign language at different proficiency levels: the case of environmental protection vocabulary

Authors

How to cite

GOST Nikitskaya N. A., Cherkashina E. L. Adaptation of authentic texts by means of generative artificial intelligence for teaching Russian as a foreign language at different proficiency levels: the case of environmental protection vocabulary // Environmental Management Issues. 2026. Vol. 5. No. 4. P. 158-169. DOI: 10.25726/u4328-6783-2990-v
APA Nikitskaya, N. A. & Cherkashina, E. L. (2026). Adaptation of authentic texts by means of generative artificial intelligence for teaching Russian as a foreign language at different proficiency levels: the case of environmental protection vocabulary. Environmental Management Issues, 5(4), 158-169. https://doi.org/10.25726/u4328-6783-2990-v

Abstract

Teachers of Russian as a foreign language increasingly use generative language models to adapt authentic texts, but it is unclear whether such adaptations match the requested level and keep subject-specific vocabulary. The study measures how environmental protection texts change when a model adapts them for levels A2, B1 and B2. Twelve authentic texts of three genres (Federal Law No. 7-FZ, Government news releases, Russian Wikipedia) were adapted by Claude Opus 5.5 and Claude Haiku 5.5 from one teacher prompt in isolated sessions (72 adaptations). We measured sentence and word length, the share of tokens beyond frequency thresholds matching the lexical minima in the Lyashevskaya–Sharov dictionary, and retention of lemmas from a term inventory drawn from the glossaries of Federal Laws No. 7-FZ, No. 89-FZ and No. 33-FZ; levels were compared by the exact Wilcoxon test. With Opus, sentence length dropped from 24.6 words to 15.7, 13.1 and 10.6 words at B2, B1 and A2, while the share of tokens beyond 1,300 most frequent lemmas differed by only 5.3 percentage points between A2 and B2 and repeated the difficulty ranking of the sources (rho ≥ 0.90). No A2 or B1 adaptation reached 5% off-level tokens. Specialised terms were retained in 94-100% of cases counting derivational pairs, while at A2 the models paraphrased the general components of legal term formulas. The model differentiates levels mainly through syntax; vocabulary control for A2 and B1 needs an explicit request or a frequency-profile check.

Keywords

Russian as a foreign language generative artificial intelligence authentic text adaptation proficiency levels environmental protection vocabulary text frequency profile term explanation

Funding

The authors did not declare any external funding for this research.

References

Gafiyatova E., Galyavieva L., Solnyshkina M. Lexical predictors of complexity of educational texts in Russian as a foreign language // Philology and Culture. 2023. no. 2(72), pp. 33-44.

Elnikova S.I. Artificial intelligence in the system of teaching Russian as a foreign language and assessing the level of proficiency in Russian as a foreign language // Russian Language Abroad. 2020. no. 2 (279), pp. 20-26.

Ilyin D.N. Features of presentation and analysis of units of the semantic field of combustion in a foreign language audience // Russian Test: Theory and Practice. 2024. vol. 10, no. 2, pp. 130-137.

Laposhina A.N., Lebedeva M.Yu. Textometer: an online tool for determining the level of complexity of a text in Russian as a foreign language // Rusistika. 2021. vol. 19, no. 3, pp. 331-345.

Norkina M., Alekseeva S., Chernova D., Kharchevnik M. Corpus of sentences for learners of Russian as a foreign language: the influence of universal parameters on lexical access in a non-native language // Russian Linguistics. 2024. vol. 48, no. 1, pp. 1-20.

Sysoev P.V., Filatov E.M. Artificial intelligence technologies in teaching Russian as a foreign language // Rusistika. 2024. vol. 22, no. 2, pp. 300-317.

Khokhlova M.V. Ecological discourse in the system of institutional discourses (to the formulation of the problem) // Bulletin of Moscow City Pedagogical University. Series: Philology. Theory of Language. Language Education. 2019. no. 2 (34), pp. 101-108.

Huang C., Slozhellikina Yu.V. The use of artificial intelligence in teaching Russian as a foreign language: adaptation of texts of scientific style // Education Management: Theory and Practice. 2025. vol. 15, no. 12-1, pp. 12-21.

Blinova O., Tarasov N. A hybrid model of complexity estimation: evidence from Russian legal texts // Frontiers in Artificial Intelligence. 2022. Vol. 5. pp. 1-14.

Kalajdžisalihović N. Readability scores and content in simplification of authentic text // ExELL. 2026. Vol. 13. № 2. pp. 181–191. DOI: 10.51558/2303-4858.2025.13.2.181.

Pushkareva E.A., Korzhenevskaya A.Y., Zdorova N.S., Lytayeva M.A., Dragoy O.V. Features of scientific text perception: an eye-tracking study // Education Issues. 2026. № 1. pp. 185-213.

Talalakina E., Brown T., Kamrotov M. Lexical profile of L2 Russian textbooks // Russian Language Journal. 2019. Vol. 69. pp. 195-212.

Watanabe M. ChatGPT, can you make this text easier?: investigating the utility of LLMs in easing text readability // International Journal of Artificial Intelligence in Education. 2025. Vol. 35. № 6. pp. 3922-3942.

Webb S., Pellicer-Sánchez A., Wang A. Lexical coverage and reading comprehension revisited // Reading in a Foreign Language. 2025. Vol. 37. № 1. pp. 1-21.

Wu J., Lin Y. ChatGPT for text simplification in Chinese as a second language textbooks: a comparative study with expert adaptations // International Journal of Applied Linguistics. 2026. pp. 1-12.

Metrics

4 views
0 downloads
Want to publish with us?
Submit an article