![]() |
Lee Jun-ho |
Convenience will favor those on the winning side; however, misery will haunt those on the losing side. This is a fact of life that nobody wants to swallow.
However, the technological revolution will continue to zero in on a new service area. Once any profession is pointed out as a target, it will be challenging for any industry to counter against the blast.
The case in point is the translation industry, which has been forcibly chosen as the next mission for artificial intelligence to clear. Google launched its Neural-network Machine Translation (NMT) service in 2016. The accuracy and fluency were better than its previous machine translation service. The media started to create hyperbole about machine translation, and those who have never cared about the job of translation started to talk about translation and its future.
Of course, it is understandable that lay people blindly believe that human translation can be replaced by machine translation in the future. However, what matters is that experts in translation studies and practice do not believe so. Who is right and who is wrong?
To answer this question, it is necessary to present a series of solid facts and convincing arguments about the current status of machine translation. Then, it would be helpful for readers to make an informed prediction rather than a biased assumption.
As a first step, this commentary will introduce general and social considerations pertaining to machine translation. Later, machine translation will be further analyzed based on feedback from translation practice and technology experts.
One mistake people make when discussing translation is to focus only on transfer of superficial meaning from one language to another. However, translation is a part of many social activities and can have social ramifications. For example, online translation services are provided free, but no one can be 100 percent certain that the service provider does not read, utilize or store the text data users input. Because of such security concerns, a few companies have already started encouraging employees not to use online translation services for any confidential information.
Translation is another form of linguistic activity, and any linguistic activity can have a direct impact on human behaviors.
This important principle seems to be often forgotten as too much attention is given only to technical integrity. To be more specific, a mistranslation can cause significant consequences in a high-stakes environment such as legal and medical communications.
No one is willing to undergo a surgery solely relying on machine translation, however well it may turn out. This means there must be someone, not something, who can verify whether the machine translation is correct and will not harm users. That is why machine translation will have limited areas of application as long as an emotional barrier exists in human hearts.
Lee Jun-ho is a professional conference interpreter specialized in information technology. He has been teaching translation at a graduate school and other universities for seven years, completed PhD coursework and studies intensively about the feasibility of machine translation. He can be reached at cuefit@gmail.com.