What role will the machines take in translation jobs?
Published on 6 December 2021 by Alice Chan
“Are the machines Coming for Your Job?” the Times (Ames, 2020) and New Yorker (Lepore, 2019), the two representative publications, asked similar questions in their headlines and brought out people’s concerns about the future of work and the rise of artificial intelligence (AI).
There is no doubt that machine translation becomes increasingly important to the translation industry. Machine translation does not only improve the speed of translation far beyond humans, but the translation outcome has also improved to the similar high level as human translators over these years. Moreover, machine translation can highly enhance translation performance through the use of computer-aided solutions such as translation memory and customised glossary, so as to improve productivity and revenue growth.
Artificial intelligence (AI) technology plays an important role in training machine translation systems and enhancing machine learning by replicating and simulating human intellect in machines. However, it must be admitted that AI translation system cannot operate independently without human support because it requires mass volume of useful data to support. Those data are based on the accumulation of many rounds of practice by human translators in their actual work.
First of all, language is the expression projected by our ever-changing lives and cultures. As languages continue to adapt and evolve, the system needs to be keep updating with the latest and reliable data from translators or linguists. Then, it takes effort and time to filter the noise and preprocess the real-life data used to train the artificial intelligence system to create more satisfactory results. Therefore, IT experts, programmers and researchers play an essential role in developing and improving machine translation.
The application of AI technology in machine translation has made significant leaps and bounds. However, there is no guarantee that machine translation can create a perfect translation, especially when it comes to selecting words with multiple meanings, which sometimes still relies on human touch. If users input new terminology or rare jargons which does not exist in the original training data in the original text, the machine translation system will not be able to create new translations on its own and may lead to mistranslation. AI does not only lack the innate creativity of human beings, but also requires high costs in maintenance and training. These problems cannot be solved at this stage and can only be supplemented by manpower.
Post-editing by human vetters or reviewers as the final gatekeeping is crucial to polish the machine translation afterwards. As language-related tasks are undergone depending on the evolution of natural utterances, translators and linguists need to ideate and produce a specialised and customised translation which fits the style and tone of specific genres and sectors as well as the needs of the target audience. The role of machine translation is not to replace human translators, but to act as a competent assistant to human beings and improve their work efficiency and translation quality.
Each translation is different in usage of words and sentence structure. This is the charm of language-related work, creativity and craftsmanship. At the same time, machines can learn from the previous and ongoing human experience, simulating human language and giving humans a helping hand.
Lepore, Jill. (February 25, 2019). Are Robots Competing for Your Job? The New Yorker.
Ames, Jonathan. (5 November, 2020). Are the robots coming for your job? The Times.