Are algorithms translator-friendly? Some considerations on the impact of neural machine translation on translator working conditions - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Parallèles Année : 2023

Are algorithms translator-friendly? Some considerations on the impact of neural machine translation on translator working conditions

L'algorithme sert-il les traducteurs ? Conditions et contexte de travail avec les outils de traduction neuronale

Claire Larsonneur

Résumé

Literary translation is widely seen as incompatible with machine translation, which is most effective on technical content. Yet machine translation tools can in fact be harnessed for editorial content and may improve the ergonomics of the translation process: by offering cognitive support and customization, they can potentially reduce working hours and physical strain. Like all technological innovations, machine translation has the potential to impact the socio-economic conditions of the translation profession. NMT is efficient and less costly than CAT software; however, it may disempower translators, and data exploitation is rife. The translation market appears to be shifting into the hands of a few powerful Internet companies with the financial means to develop language models. As a result, due to changes in online practices, the translation market is breaking down into distinct markets for multilingual content management, APIs and editorial publishing. This raises ethical issues and calls for new forms of regulation.
Fichier non déposé

Dates et versions

hal-04395321 , version 1 (16-01-2024)

Identifiants

Citer

Claire Larsonneur. Are algorithms translator-friendly? Some considerations on the impact of neural machine translation on translator working conditions. Parallèles, 2023, 35 (2), ⟨10.17462/para.2023.02.10⟩. ⟨hal-04395321⟩
7 Consultations
2 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More