Since coming to Japan two years ago as a Japanese-to-English translator, I’ve been wanting to start a blog for the four other people on this planet who care about the extraordinarily specific field of programming’s use in translation (although I will talk about other stuff too, probably). AI has come up frequently in the translation field as this lingering threat with people in many different fields having an array of different opinions. To some, it’s simply a useful tool that translators will need to adapt to in order to remain competitive in the field. To others, it’s the proverbial sword of Damocles suspended above the neck of every translator who have invested too much time in the field to run away. There’s much debate on who will be the first to go as well — many of the arguments in favor of which fields are untouchable suspiciously coming from people who work in that field.

Creative translation, such as for anime, manga, novels, etc. is protected by the human touch required to acknowledge and work around cultural differences. Localization of text is more than just replacing words of one language for another — concepts that are commonplace or obvious to one culture can be foreign to another, and this kind of deep understanding of culture is something only a human brain can understand at this point in time. This is probably the most convincing argument for a field that is “safe from AI”, yet in practice it seems that companies are ignoring this and attempting to push it regardless more than any other field of translation. These translations are always universally panned their audiences. It doesn’t require any Japanese knowledge to see that the results from AI translation come out poor and at times borderline incomprehensible. Unfortunately, some translation companies have historically always cut corners in order to cut down on the costs of translation at the expense of translation quality. This means that the question isn’t really if AI can do as good of a job as human translators, but can it get to the point where the quality is acceptable enough that it is more profitable to the company—the point where the text isn’t so bad that a general audience can watch it and not notice. If AI can be used in part of this process to cut down costs, such as translation with post-editing, it will become a more enticing option to translation companies.

Other fields, such as legal / patent translations, business translation, medical translation, etc. don’t have this same obstacle of needing a human to understand the deeper cultural nuances of the text, but the stakes of the translation being significantly higher prohibits the use of AI. I’ll talk more about this in another post.

AI, for better or worse, will affect the field of translation. Whether you wish to adopt it or protest against its use, it’s something every translator is going to have to remained informed about. I hope this blog will succeed in keeping others in lockstep as AI progresses.