Will Machines Translate Languages Better Than Humans Someday?

I believe machine learning software such as Artificial Intelligence will replace our need to translate languages in the same way calculators have replaced our need to do basic math. Our interest in understanding languages to help people of different cultures communicate will disappear to some extent and will change as a whole in the future. As the enterprise of Artificial Intelligence increases in sophistication, it will provide large amounts of data that will communicate particularly in widely spoken languages including English, which is potentially useful in a lot of different scenarios.

As a graduate in Linguistics, I think fully translating and interpreting general languages by machine are within reach. I feel confident in machine translation and its abilities to replace people in translating simple constructions into English as closely as possible in this day and age. However, the main thing is its struggles with context, which there are many with unusual constructions. With Artificial Intelligence, companies like Google will have at least an idea about computing dialect or linguistic variation, as well as recognition of context, register, tone, and pitch in a spoken and written language.

In my experience, translators simply can’t be expected to translate anything more than simple phrases, but for the purposes of understanding context without giving the phrase, the translators will eventually fill in many lexical and pragmatic gaps to have a chance at being accurate. The speed at which any individual human being is capable of assimilating new words and phrases can’t be matched by the power of predicting precise answers over a period of time. As ironic as it is, machines will translate the way we interpret human languages and communication.

What machine translation will do in the next decade is potentially raise very interesting concepts about understanding translation and human perception of our own thought processes. What we don’t know about our brain’s ability to learn is within the framework (process of analysing and producing) of machine based learning.  It’s still quite a lot for artificial intelligence to learn in a short period of time, but it has the capabilities to become more sophisticated. How it will achieve it can reveal some interesting problems to the public perception of what is possible.

A misconception that I hear is it will only benefit anglophone speakers from translation. This outlook is bleak to me because it’s not as suggested to be anything more than a way to connect to common cultures. The former argument of language (from a linguistic standpoint)  is that it’s suppose to be universal, which means that all languages can be traced back to a common ancestor. Language is more than being an organic structure to a particular group of speakers.

The knowledge economy is going to thrive in a few years when machines catch up and that will affect the perceived value of our education and skills. I can cope with that kind of change because it might probably affect the way we communicate knowledge with each other. As Artificial Intelligence incorporates into machine translation, there will be some not fully answered philosophical questions about the value of one language over another. As I have pointed out earlier, the machine translation will be based on a framework. Within that framework, a simulation of the human experience is much written and recorded in languages around the world.

Because a company like Google has sophisticated technology and tons of recorded data, their starting point is to translate from a machine-based code (0 and 1 in the binary system of ‘on’ versus ‘off’, if you will) and use it as the lingua franca language in a 3 way translation. The method resorts to English when data is lacking, since it’s the most widely translated language. This puts English at the top of a global language hierarchy, at least in translation. Since Google have been a wildly successful company, things might be different as we continue to use their databases for searches, videos (YouTube), and advertisement.  They might have the English-speaking world widely connected around the quality of and access to goods and services coming out of their dominance on the Internet, but the machine’s “language” will also interpret and translate two non-English languages with the same level of accuracy as English.

At the lower end of the scale, adults and children will get ahead without learning a new language. Because they won’t struggle with translations, they’ll learn enough in school to do well in the job market. Their untranslated words would no doubt have blunders that machine translation will easily fix. Or better yet, the availability of information and products will provide access to cutting-edge research or decent opportunities to travel abroad. On the higher end of the scale, competition in an open global market will affect any language translation as being worthy on a path to greater equality. Corporations who buy ads on Google will affect it’s worth by translating their instructions properly into smaller languages or to languages that are spoken within smaller populations of people. Is it worthy to display the nutritional information at McDonald’s in Latin?  Maybe yes, maybe not, but you can guarantee that the non-English speakers will have the same opportunities as the native English speakers.

As more examples of what Artificial Intelligence can do to the world come to surface, corporations will make decisions to spend money in ways that might put cultures in direct competition with profit margins. That puts a literal and figurative price on languages and cultures, which is the opposite of what language diversity and language translation is all about. The problem, as represented by the incarnation of machine translation, is that it represents a world in which we presume to assign value to what we don’t know. As a result, we assign different values to different languages, creating a conceptual hierarchy that has real-life ramifications. As I described above, this might cause competition, and all of that which it entails, where the line between language and culture is so thin that it’s up for debate.

The question in my mind at this point is when the time machine translators continue to improve on the spirit in which it operates, will it go about it in the right direction? That’s why I think the engineers at Google work diligently on Artificial Intelligence and machine based translation. Because Google got enough money to employ the best, it’s only a matter of time before most of us can conceptualize the tricky concept of translating the world’s languages within its framework with accuracy and precision beyond human interaction. As the number of people committed to the endeavor increases, we can only assume that machines will translate languages in a way we never can.  

Recommend0 recommendationsPublished in Foreign Languages, Technology