This paper discusses neural machine translation (NMT), a new paradigm in the MT field, comparing the quality of NMT systems with statistical MT by describing three studies using automatic and human evaluation methods. Automatic evaluation results presented for NMT are very promising, however human evaluations show mixed results. We report increases in fluency but inconsistent results for adequacy and post-editing effort. NMT undoubtedly represents a step forward for the MT field, but one that the community should be careful not to oversell.

The profession of translation is undergoing enormous change, expedited by the global recession that began in 2007–2008. Government policies and an intensive focus on cost has resulted in a large scale move to freelance or contingent work, leaving the worker in a precarious position with regard to rights and undermining his or her agency. This shift is exemplified particularly by the vendor model widespread in specialised translation work. Related to the downward pressure on costs and productivity is the technologisation of translation, with translation tools becoming a necessity and new use cases being found for post-edited and raw machine translation. Despite the recession, continued growth has been reported for the language industry, and the outlook for employment in translation is positive. This paper looks at the background to the economic and technological changes to translation, attempts to put them into a wider context, and looks to the options available to translators to maximise their agency within the ‘global value chain’. Translators have little option but to embrace new competences, but also need to focus on their expertise to maximise leverage and agency.

In the context of recent improvements in the quality of machine translation (MT) output and new use cases being found for that output, this article reports on an experiment using statistical and neural MT systems to translate literature. Six professional translators with experience of literary translation produced English-to-Catalan translations under three conditions: translation from scratch, neural MT post-editing, and statistical MT post-editing. They provided feedback before and after the translation via questionnaires and interviews. While all participants prefer to translate from scratch, mostly due to the freedom to be creative without the constraints of segment-level segmentation, those with less experience find the MT suggestions useful.

Machine translation is currently undergoing a paradigm shift from statistical to neural network models. Neural machine translation (NMT) is difficult to conceptualise for translation students, especially without context. This article describes a short in-class evaluation exercise to compare statistical and neural MT, including details of student results and follow-on discussions. As part of this exercise, students carry out evaluations of two types of MT output using three translation quality assurance (TQA) metrics: adequacy, post-editing productivity, and a simple error taxonomy. In this way, the exercise introduces NMT, TQA, and post-editing. In our module, a more detailed explanation of NMT followed the evaluation. The rise of NMT has been accompanied by a good deal of media hyperbole about neural networks and machine learning, some of which has suggested that several professions, including translation, may be under threat. This evaluation exercise is intended to empower the students, and help them understand the strengths and weaknesses of this new technology. Students’ findings using several language pairs mirror those from published research, such as improved fluency and word order in NMT output, with some unpredictable problems of omission and mistranslation.

Translators have worked with the assistance of computers for many years, usually translating whole texts, divided into segments but in sequential order. In order to maximise efficiency and inspired by similar moves in the tech industry and predictions for Industry 4.0, large translation companies have begun to break tasks down into smaller chunks and to rigidly define and monitor translation processes. This is particularly true of platform-mediated work, highly collaborative workflows, and multimedia work that requires near-live turnaround times. This article considers such workflows in the context of measures of job satisfaction and discussion of sustainable work systems, proposing that companies prioritise long-term returns and attempt to balance the needs of all stakeholders in a translation process. Translators and translator trainers also have a role to play in achieving this balance.