Nowadays, everybody is talking about Machine Translation.
Some people say it will soon replace human translators.
Some say it is only good for creating funny, yet rather embarrassing mistranslations.
Some say it is just a tool and nothing else.
- Which of the above statements is closer to the truth?
- How does it work?
- As a business owner, should I use Machine Translation services?
- Which are its main benefits? What can go wrong?
Below, you will find the answers to many of the most common questions asked around this fancy technology that is unquestionably reshaping the language industry as well as the information distribution model adopted by businesses all over the world.
1) What is Machine Translation?
Even if the underlying technology is incredibly (and increasingly) complex, to put it simply:
Machine Translation (MT) is a technology that helps process an original text in a given language — what we call source language — to produce a translated version — into what we call a target language.
The process is mainly based in a mix of language rules (syntax, grammar), statistical data (i.e. which words tend to be found near other words) and a text corpus made up of millions of words from real documents and files that are used by the MT engine to check and predict the best possible match for each sentence. Some MT providers provide the ability to use either generic engines (created from all sorts of texts and documents) or specific to certain fields such as engineering, legal, medical, etc.
Machine translation has significantly evolved over time and it has recently conquered a lot of headlines, which may lead to thinking the technology is relatively new; however, the history of MT goes back to the 1930’s (yes, it’s not a typo, I said nineteen thirties), when Peter Troyanskii, a Russian scientist, devised a machine that could mix and match groups of words printed in cards using four different languages (including Esperanto!). Not the most accurate system, of course, but that opened the door to countless possibilities on the field of translation automation processes.
We have obviously walked a long path since then, and the level of accuracy and possible applications of the technology are increasing by the day. Nevertheless, is it suitable for every purpose?
2) Should I use MT to provide my business content in multiple languages?
The nature of your business determines the nature of your texts. Your company works with specific services, markets, competitors and digital spaces where your communication takes place. This defines the language used in your blog posts, user manuals, internal and client-oriented newsletters, marketing material, contractual documentation, and regulatory sheets, and each of these feature a different style and purpose.
However, the translation process should not stop there by any means, and there are good reasons to follow this piece of advice.
If your company creates large pieces of information that need to be translated into different languages, MT can help speed up the process and provide a draft version of your texts.
3) Will the quality of my translations drop if I use Machine Translation?
As you may have guessed by now, Machine Translation is a tool, not a substitute. As such, it does not understand the meaning of the text being translated, nor is it capable of catching the nuances, the intention, the full context and the style of the original writing.
We are not talking about literary translation with hidden meanings:
Imagine an electric diagram or a software application where the term “Open” needs to be translated into French, Spanish and Italian.
The content owner has provided the text using an Excel spreadsheet, and there is not enough context to determine whether that “Open” is meant to be a command (so it should be translated using a verb in infinitive form), an instruction (so it should use imperative), a status indicator (so it should be translated as an adjective) or if it is part of a longer sequence of terms and needs to be translated together with the previous or the following cells.
Only a human translator can check those references and choose the correct form to be used in each case. In this case, when a real person checks and eventually adjusts the output from a MT engine, we call it Post-Editing or MTPE.
4) What is Machine Translation Post-Editing (MTPE)?
Good! The Machine Translation engine has already pre-processed your documents, creating a set of translated files that are ready to be used! Well, not quite, as we’ve just seen.
As you will soon find out, MT does a pretty decent job processing many types of text, preferably using not-too-complex sentence structures and a straightforward style. However, remember that a MT engine does NOT understand the text: it simply processes words following a complex pattern, but cannot decide whether a word with several possible interpretations has been used with a specific intention.
Once again, only human translators can understand those nuances and make an educated decision about the best fit for each translated word. Leaving MT text untouched leads to funny mistranslations in the best case, but it might also lead to catastrophic errors (imagine what can happen if a legally binding term is mistranslated in a contract, or in a manufacturing scheme for aeronautical devices). The human step is not simply recommended: it is essential to guarantee the quality AND safety of the MT output.
5) In which contexts is MT typically used?
This also depends on the amount of time and money you are willing to spend on refining the rough output provided by the MT engine.
Theoretically, MT can be used in any kind of text, provided that the source material is correctly written and has a clear structure and no major mistakes that may prevent the machine -and any reader- from understanding the content.
However, there are some industries where MT works better thanks to documents featuring a more straightforward layout that helps a machine read the information more efficiently. Technical documentation, user guides with large parts of predictable text structures and medical reports are a good fit, but the scope of MT is becoming wider every day, and it can be applied to media or legal translations with an increasing degree of accuracy. Generally speaking, the less creative a text is, the higher the chances that MT produces a decent raw output.
In any case, if you are not sure about whether MT can be used to process your information, simply ask us. If the effort involved in post-processing the MT output is bigger than using purely human translation, better keep the old school ways!
6) Will my business content become publicly available if I use it?
Short answer: Under normal circumstances, no.
Long answer: Machine Translation engines (that is, the tools that process your data) running on private servers owned by the company providing the MT service (with very few exceptions, this is not the language service company you hired to translate your texts). Language service providers (LSP) rely on 3rd party MT services to carry out this step, and then pick the pre-translated files to add the human touch to your text.
Both the language service provider (for example, Ampere Translations) and the MT provider (Google or Microsoft, just to name two popular services) are bound by strict regulations concerning how and where your data is stored. For companies operating within the EU (or engaging in commercial activities with EU companies), GDPR provides a safe framework to ensure that your data can be traced and used only for the intended scope —in this case, translation. Every step of the chain (from project management to translation suppliers and project managers at the LSP) must be compliant with GDPR requirements.
Of course, if you have any specific questions, for example regarding the location of the servers where your information is stored and where the MT process takes place, you have the right to ask, but usually you do not need to worry about it.
7) Does — or will — Machine Translation replace human translators?
Short answer: Never.
Long answer: Back in the early 90’s, there were already some authoritative voices stating that MT was this close to achieving technical perfection (the long-sought FAHQMT, that is, fully-automated, human-quality machine translation).
Since then, many other voices have been sharing and shouting the same apocalyptic prediction relentlessly. However, human translation is still here, and it is meant to stay.
Leave aside the technicalities of the process behind MT learning, which leads to consistently improved versions of this technology. No matter how refined the tech becomes, there are several inherent aspects of communication that no machine can fully understand: intention, style and communicative context. Failing to reproduce them inevitably result in hollow or plainly wrong translations. We do not need to remind you that you should never underestimate the negative impact of a mistranslation in your business.
In order to ensure that these ingredients are properly added to the translated text, a human touch is required.
You might argue that there are currently AI solutions used in several industries that can produce (create!) automated texts for a number of purposes. Of course, this kind of robotic output is the perfect candidate for raw MT: however, the quality requirements are completely different for such disposable, quick-consumption type of information than for text that is meant to stay around for a long time or targeted at a demanding audience.
If quality is what your business seeks, then human translation —perhaps combined with Machine Translation as we’ve explored in this post— are your best choice. You simply need to understand the purpose of your text and act accordingly. Feel free to ask!