One major limitation of machine translation is its reliance on direct word-to-word translation, which often fails when dealing with legal terms that require contextual adaptation.
My understanding is that legal discourse uses relatively "objective" language to avert misinterpretation by legal bodies. As such, MT may be a safer way of translating it, knowing that legal discouse is almost void of metaphor, metonymy, idioms, irony, and proverbs. Having myself experience with translating legal contracts, I was able to notice this literalness and repetition in my own translations to avoid all sorts of ambiguity.
According to my experience, MT alghoritms should be changed in order to streamline MT translation processes. Some scholars posit that corpus consultation can be integrated in MT processes. I tried myself to integrate MT output with corpus evidence and it worked satisfactorily. However, it was a very time consuming activity. For this reason, some scholars have suggested APE (Automated Post-Editing) systems which automatically correct MT output (do Carmo et al. 2020). Such APE systems could learn from parallel corpora (Chatterjee et al. 2017; Negri et al. 2018).
References:
Chatterjee, Rajen, Gebremedhen Gebremelak, Matteo Negri, and Marco Turchi. 2017. Online automatic post-editing for MT in a multi-domain translation environment. In Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics:
Volume 1, Long Papers, 525–535. Valencia, Spain.
do Carmo, Félix, Dimitar Shterionov, Joss Moorkens, Joachim Wagner, Murhaf Hossari, Eric Paquin, Dag Schmidtke, Declan Groves, and Andy Way. 2020. A review of the state-of-the-art in automatic post-editing. Machine Translation 35: 101–143.
Escribe, Marie, and Mitkov Ruslan. 2023. Applying Incremental Learning to Post-editing Systems: Towards Online Adaptation for Automatic Post-editing Models. In Jun Pan and Sara Laviosa (Eds), Corpora and Translation Education, Advances and Challenges. Singapore: Springer Nature Singapore,35-62.
Giampieri, Patrizia. 2024. AI and the BoLC: streamlining legal translation. Comparative Legilinguistics, 58: 67-90.
Negri, Matteo, Marco Turchi, Nicola Bertoldi, and Marcello Federico. 2018a. Online neural automatic post-editing for neural machine translation. In Proceedings of the 5th Italian Conference on Computational Linguistics (CLIC-IT 2018), 288–293. Torino, Italy.
Negri, Matteo, Marco Turchi, Rajen Chatterjee, and Nicola Bertoldi. 2018b. eSCAPE: A large-scale synthetic corpus for automatic post-editing. In Proceedings of the 11th International Conference on Language Resources and Evaluation (LREC 2018), 24–30. Miyazaki, Japan.