Practical Dilemmas, Ethical Risks and Human-Machine Collaboration Pathways for AI-Empowered Legal Translation
DOI:
https://doi.org/10.54097/1abf1f81Keywords:
Legal English, Legal Translation, Artificial Intelligence, Human-machine Collaboration, Machine TranslationAbstract
The rapid iteration of generative artificial intelligence (AI) and neural machine translation has brought efficiency innovation to legal translation practice, while posing new challenges to the conversion of legal English terminology, transmission of concepts across legal system and translation quality control. Different from general text translation, legal translation features dual attributes of linguistic transformation and transplantation of legal cultural connotations. AI translation has inherent limitations in the accurate matching of legal terms, disambiguation of legal senses of polysemous words, reproduction of legal stylistic features, and reconstruction of legal logic. It also gives rise to multiple ethical risks including translation distortion, data bias, algorithmic black-box effects, technological alienation and blurred accountability. Against the backdrop of global expansion of cross-border commercial transactions and international judicial cooperation, indigenous Chinese rule-of-law discourse requires reliable translation output to eliminate cross-jurisdictional misunderstandings. Nevertheless, off-the-shelf large language model outputs frequently produce superficially fluent yet legally inaccurate renderings, a phenomenon known as model hallucination, which aggravates risks in high‑stakes legal scenarios. Drawing on empirical cases of legal terminology translation, this paper figures out the strengths and practical bottlenecks of artificial intelligence applied to legal translation, analyzes the root causes of translation dilemmas, and constructs a human-machine collaborative workflow adapted to legal translation scenarios. Optimization suggestions are put forward from the dimensions of talent cultivation, industrial norms, technical supervision and terminology database construction, so as to facilitate the accurate cross‑cultural communication of indigenous Chinese rule-of-law concepts.
Downloads
References
[1] Gu, W. C. (2011). On the role of legal translation talents in foreign-related economy. Hebei Academic Journal, 31(4), 179–182.
[2] Wang, H. S. (2025). Ethical risks of generative-AI translation technology in legal translation. Journal of Gansu University of Political Science and Law, (01), 1–12.
[3] Fu, J. M. (2025). Applied-translation research in the age of artificial intelligence: Crisis or opportunity. Foreign Language Teaching, (02), 82–87.
[4] Albi, A. B., & Martínez-Carrasco, R. (2019). Future-proofing legal translation: A paradigm shift for an exponential era. In I. Simons & M. Kristiansen (Eds.), Legal translation: Current issues and challenges in research, methods and applications (pp. 187–206). Printed in Germany.
[5] Killman, J. (2023). Machine translation and legal terminology: Data-driven approaches to contextual accuracy. In Handbook of terminology (p. 494). John Benjamins Publishing Company. DOI: https://doi.org/10.1075/hot.3.mac2
[6] Zhang, F. L. (2020). On machine translation technology in legal translation. Technology Enhanced Foreign Language Education, (01), 53–58.
[7] Wu, C. H., & Zou, J. T. (2025). Challenges and optimization of machine translation for legal terms: Term-generation mechanisms and human-machine-collaboration strategies. Shanghai Journal of Translators, (06), 56–61.
[8] Wang, K. Y., & Cheng, S. (2025). A study on the effectiveness of ChatGPT in legislative-text translation and quality evaluation. Journal of Zhejiang Ocean University (Humanities & Social Sciences Edition), (01), 86–95.
[9] Cui, Q. L., & Lei, X. F. (2016). Human-machine interactive translation strategies based on text stratification. Contemporary Foreign Languages Studies, (03), 44–48.
[10] Zhang, L. (2025). Research on post-editing strategies for AI-assisted legal text translation. Chinese Translators Journal, (05), 94–101.
[11] Zhang, F. L. (2023). Research on the disciplinary planning of legal foreign language studies. Technology Enhanced Foreign Language Education, (03), 3–9, 113.
[12] Wang, H. S. (2023). Research on translation technology. Foreign Language Teaching and Research Press.
[13] Feng, Z. W. (2011). Introduction to modern terminology (Enlarged edition). The Commercial Press.
[14] Nida, E. A. (2004). Toward a science of translation. Shanghai Foreign Language Education Press.
[15] Lai, J. Q., Gan, W. S., Wu, J. Y., et al. (2023). Large language models in law: A survey. arXiv preprint arXiv:2312.03718. https://doi.org/10.48550/arXiv.2312.03718.
[16] Dahl, M., Magesh, V., Suzgun, M., & Ho, D. E. (2024). Large legal fictions: Profiling legal hallucinations in large language models. Journal of Legal Analysis, 16(1), 64–93. https://doi.org/10.1093/jla/laae004. DOI: https://doi.org/10.1093/jla/laae003
[17] Song, L. Y. (2024). Research on the digital-humanities transformation of legal translation: Centering on thematic databases and ChatGPT. Foreign Language Research, (2), 51–57.
[18] Zhang, F. L., & Qu, X. (2021). Exploring the construction of technical ethics for legal translation. Foreign Languages in China, 18(6), 17–22.
[19] Editorial Group of Terminology of the Civil Code. (2021). Terminology of the Civil Code. People’s Publishing House.
[20] Bajcic, M. (2024). Applying large language models in legal translation: The state-of-the-art. International Journal of Law and Language. DOI: https://doi.org/10.2139/ssrn.5040587
[21] Meng, L. Y., Liu, M. L., Wang, H., et al. (2025). Building from scratch: A multi-agent framework with human-in-the-loop for multilingual legal terminology mapping. Artificial Intelligence and Law. DOI: https://doi.org/10.1007/s10506-025-09490-6
[22] Tiersma, P. M. (1999). Legal language. University of Chicago Press. DOI: https://doi.org/10.7208/chicago/9780226854809.001.0001
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Journal of Education and Educational Research
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
