Maschinelle und menschliche Übersetzung im Vergleich am Beispiel des Entwurfs des ersten ungarischen Zivilgesetzbuches (Deutsch – Slowakisch)

Abstract

Neural machine translation enables fast and automated translation of large volumes of text, which, in many cases, can be more efficient than translation by human translators. However, for specific communication purposes, the quality of machine translation remains a challenge. This also applies to legal communication, as legal texts require a high level of accuracy, precision and expertise that cannot always be provided by machine translation systems. This article compares the translations of an inheritance law text from German into Slovak produced by four translation tools (DeepL, Google Translate, eTranslation, ChatGPT) with human translation. The study focusses on the accuracy of the translation of the relevant inheritance law terminology and the possible causes of incorrect and therefore unacceptable translation solutions. The results of the study show that, despite the progress made, neural machine translation is not yet sufficiently mature for a translation in accordance with legal norms.