Revolt or Submit: The Future of Translators in MT-based Localization
Wednesday, 27 March, 2019
Most of the current discourse on MT focuses on the technology itself and the implications it carries for translation costs and effectiveness. By now we take it for granted that MT is here to stay and that agencies will have to adapt accordingly. But little attention is devoted to the professional translators still providing the majority of the necessary human input, albeit more and more in the form of post-editing rather than translation. While translators are typically invisible, both as the source of the large corpora used for machine-learning, and as active participants turning raw MT output into full-quality content, the industry is still largely dependent on their expertise and linguistic work. This talk aims to address the new role expected of translators, the changes in skills this requires, and the level of acceptance these changes are met with from the translators’ side.
Firstly, post-editing as a new workflow will be discussed based on existing literature and personal experience. Secondly, the talk will focus on the acceptability of these changes to translators. Grounded in the results of a survey carried out among freelance translators through online translators’ fora (proz.com, translatorscafe.com), the presentation will map the main advantages and reservations translators express with regard to working in an MT-dominated environment. Finally, possible suggestions will be considered regarding scenarios feasible for all parties involved in MT post-editing, focusing on user experience and the reasons for potential resistance.
Zsófia Gregor is Linguistic team lead at UK-owned life sciences and technical specialist LSP Albion Languages. She is responsible for coordinating and overseeing all internal linguistic work. During her 5 years at the company, she has worked with a team of ten in-house linguists and a pool of over 400 external resources on a wide variety of demanding projects. Her team is accustomed to ambitious deadlines and diverse CAT tool requirements. Zsófia holds a Master’s degree in American Studies and earlier taught English, as well as being an editor at a publishing house.
Most machine translation output is fine for basic Western LSP needs, as they usually translate within European...
Machine translation is great when it works, but it fails often and in unpredictable ways. What causes bad machine...