Nowadays many people wonder if the progressive development of artificial intelligence technologies is a threat to certain professional roles. Among these there are translators, for instance. Some people say that in a few years professional translators will disappear, since their job will be simply done by machines. By extension, one might think that the same could happen with the localization of creative products, such as video games. But is that so? Let’s find out if it will ever be possible to automate localization!
A brief introduction to automated translation
Automated translation is a branch of computational linguistics and translational science. It’s a discipline that deals with the translation of texts from a natural language to another by using algorithms. In automated translation a computer translates without human intervention.
This approach to translation is very convenient because of its efficiency. One of the most well-known web apps based on automated translation is Google Translate, for example. It is centered around one type of automated translation, Rule-based Translation. Other methods are Corpus-based Translation, Context-based Translation, Statistical Machine Translation and Neural Machine Translation.
Automated translation to localize video games
But now let’s go back to the topic at hand: can we use AI to translate and localize works filled with creativity such as video games? On our blog we discussed at length about how important video game localization is. The reason is that these kinds of works often present cultural aspects which demand to be adapted for another audience. A computer would never be able to catch the cultural nuances inherent to the text. Therefore, creating completely automated video game translations is not possible.
It’s a fact that game developers and publishers largely use automated translation methods in order to localize their interactive products. But these tools are only auxiliary to their job. In fact, it is necessary that one or more professionals supervise the work of the software in a process named Quality Assurance. A computer wouldn’t understand cultural references or correctly interpret the relationship between signs and context pertaining to pragmatics. Language is always produced within a context, after all, and in automated translation this is not taken into account.
Conclusions
Completely automated localization is a very ambitious idea that will hardly become reality. Although AIs are being instructed with the goal of delivering increasingly more convincing results, human presence is still essential in granting text accuracy. A translator’s creativity while localizing is an important ingredient for the success of a video game, as shown by the transcreation technique, for instance.
