Elamur Research⌕
☰
home / papers / 2026 / Teaching linguistic prompt control for LLM based translation: A classroom approach to developing critical and responsible AI literacy

Teaching linguistic prompt control for LLM based translation: A classroom approach to developing critical and responsible AI literacy

Summary

The paper tackles a teaching problem. Translation students often prompt LLMs by unsystematic trial and error, and unsteered output shows typical "AI fingerprints", such as register-neutral, flattened style or repetitive non-finite structures. The author describes a one-semester MA seminar at Leipzig University designed to build critical, responsible AI literacy, with human translators keeping final authority.

Students first produced minimal-prompt baseline translations of an EU informational text and examined their default patterns. They then used text analysis based on Systemic Functional Linguistics (SFL) and features drawn from parallel and comparable corpora to write register-specific instructions a model can interpret. These instructions were organised into a three-part prompt structure. Students also compared commercial, institutional GDPR-compliant and local models.

According to the author's qualitative observations, the approach:

  • improved translation quality over unsteered output,
  • cut down on wasteful prompt iterations, and
  • made students more aware of each model's limits.

The study also found that models apply abstract linguistic rules unreliably and work better with concrete examples. This matters because it shows linguistic expertise remains central when working with LLMs.

Key points

  • Source text and languages: students translated an EU web text of about 1,200 words on antimicrobial resistance. English-German was the main pair, with English-French as an option, and United Nations texts were considered for other language combinations such as Arabic.
  • Prompt structure: every prompt was built from three parts: configuration instructions, the task itself, and a revision step. In the revision step, the model re-generates under constraints and can flag requirements it could not meet.
  • Tools used: ChatGPT, Microsoft Copilot (institutionally licensed), Google Gemini, Academic Cloud, Jan.AI and TranslateGemma. Local models were introduced in a workshop by Martin Czygan of Leipzig University Library.
  • Background frameworks: CO-STAR and CLEAR are cited as general prompting frameworks that need refining for specific registers.
  • Main limitation: models applied stylistic constraints inconsistently, for example by overusing participle clauses in English output. Explicit instructions with negative examples worked better, though still imperfectly. All results are preliminary and qualitative, with no quantitative metrics reported.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Linguistics & languages · Machine translation

Cite

The paper itself, not this page. BibTeX from the library's record of the work.

@inproceedings{menzel2026teaching,
  title = {{Teaching linguistic prompt control for LLM based translation: A classroom approach to developing critical and responsible AI literacy}},
  author = {Menzel, Katrin},
  booktitle = {Proceedings of the 1st International Workshop on Teaching AI-Based Translation and Technologies (TAITT 2026)},
  year = {2026},
  url = {https://aclanthology.org/2026.taitt-1.4/}
}

References

20 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Angelone, Erik. 2026. Generative AI as a facilitator of deliberate practice in translator training. In Penet, J. C., Joss Moorkens, and Masaru Yamada, editors, Teaching Translation in the Age of Generative AI: New Paradigm, New Learning?, pages 27–47. Language Science Press, Berlin. find
  2. Ayvazyan, Nune, Yu Hao, and Anthony Pym. 2024. Things to do in the translation class when technologies change: The case of generative AI. In Peng, Yuhong, Huihui Huang, and Defeng Li, editors, New Advances in Translation Technology: Applications and Pedagogy, pages 219–238. Springer, Singapore. find
  3. Ayvazyan, Nune. 2025. Human-centered pedagogies in the age of generative AI: Examining student perceptions of augmentation, progress and agency. InContext: Studies in Translation and Interculturalism, 5(1):167– 193. find
  4. Burkert, Tomáš. 2025. LLM benchmarking leaderboard: Languages and creativity tasks. RWS Blog. find
  5. Cui, Feng, Defeng Li, and Chiyuan Zhuang. 2025. Introduction: Transforming translation education through Artificial Intelligence. The Interpreter and Translator Trainer, 19(3-4):227–233. find
  6. Finkelstein, Mara, Isaac Caswell, Tobias Domhan, Jan-Thorsten Peter, Juraj Juraska, Parker Riley, Daniel Deutsch, Geza Kovacs, Cole Dilanni, Colin Cherry, Eleftheria Briakou, Elizabeth Nielsen, Jiaming Luo, Kat Black, Ryan Mullins, Sweta Agrawal, Wenda Xu, Erin Kats, Stephane Jaskiewicz, Markus Frei… find
  7. Gao, Yuan, Ruili Wang, and Feng Hou. 2023. How to design translation prompts for ChatGPT: An empirical study. arXiv. find
  8. Halliday, Michael A. K. and Christian M. I. M. Matthiessen. 2013. Halliday’s Introduction to Functional Grammar (4th ed.). Routledge, London/New York. find
  9. He, Sui. 2024. Prompting ChatGPT for translation: a comparative analysis of translation brief and persona prompts. arXiv. find
  10. Leo S. Lo. 2023. The CLEAR path: A framework for enhancing information literacy through prompt engineering. The Journal of Academic Librarianship, 49(4):1– 3. find
  11. Pym, Anthony and Yu Hao. 2024. How to Augment Language Skills: Generative AI and Machine Translation in Language Learning and Translator Training. Routledge, London/New York. find
  12. Ren Xiaobin and Wang Ruoxuan. 2025. Integrating human-AI collaboration into translation education: A comprehensive protocol for assessment, diagnosis and strategy development. PLoS One, 20(12):1–8. find
  13. Siu, Sai Cheong. 2024. Revolutionising translation with AI: Unravelling Neural Machine Translation and generative pre-trained Large Language Models. In Peng, Yuhong, Huihui Huang, and Defeng Li, editors, New Advances in Translation Technology. Applications and Pedagogy, pages 29–54. Springer, Singap… find
  14. Sizov, Fedor, Cristina España-Bonet, Josef Van Genabith, Roy Xie, and Koel Dutta Chowdhury. 2024. Analysing translation artifacts: A comparative study of LLMs, NMTs, and human translations. In Proceedings of the 9th Conference on Machine Translation, pages 1183– 1199, Miami. Association for Computat… find
  15. Spiridonov, Ilya. 2026. Best LLM for translation in 2026: A data-driven engine scoreboard. Alconost Blog. find
  16. Teo, Sheila. 2023. How I won Singapore’s GPT-4 prompt engineering competition. Towards Data Science. find
  17. Yamada, Masaru. 2023. Optimizing machine translation through prompt engineering: An investigation into ChatGPT’s customizability. In Proceedings of Machine Translation Summit XIX, Vol. 2: Users Track, pages 195–204, Macau. Asia-Pacific Association for Machine Translation. find
  18. Yamada, Masaru. 2026. Teaching translation with AI: Bridging theory and practice through prompt engineering. In Penet, J. C., Joss Moorkens, and Masaru Yamada, editors, Teaching Translation in the Age of Generative AI: New Paradigm, New Learning?, pages 87–104. Language Science Press, Berlin. find
  19. Yang, Zhishen, Tosho Hirasawa, Edison Marrese-Taylor, and Naoaki Okazaki. 2024. Large language models as manga translators: A case study. In Proceedings of the 30th Annual Meeting of the Association for Natural Language Processing, pages 2012–2017, Tokyo, Association for Natural Language Processing. find
  20. Zhang, Biao, Barry Haddow, and Alexandra Birch. 2023. Prompting large language model for machine translation: A case study. In Proceedings of the 40th International Conference on Machine Learning, pages 41092–41110, Honolulu. Journal of Machine Learning Research. find our card

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/menzel-2026-teaching-linguistic-prompt-control-llm/