What’s the real impact of machine translation on human translation? When you hear someone saying ‘I’m a translator’, you likely find yourself thinking ‘we don’t need translators, we have Google Translate! Will it really machine language translation replace humans altogether? Or has it replaced them already?
What is Machine Translation Exactly?
It’s good to clarify what is machine translation, also known as MT, before delving deeper into trying to answer these questions. The expression is pretty self-explanatory. It denotes a translation made by a machine translation software -a computer programme- for instance, Google Translate.
How does MT Works?
Language machine translation software is really easy to use, you just type in a machine translation website or a machine translation app the word or sentence you would like to translate.
Then you select the language which you want the programme to translate it into, and voilà!
You magically have the translation you were looking for. As simple as eating a piece of cake, isn’t it? Not really.
The Limits of Machine Translation Software
Although machine translation programmes had become far more reliable than when they first appeared in 2006, they still can’t pick up every little shade of language in context as professional human translators and professional subtitlers do.
And most probably they will never be able to do so. As they rely on a combination of statistics and artificial intelligence that more often than not clash with a language lexicon, which often has funny outcomes.
For example, in 2015 Google Translate confused Galician with Portuguese and translated the “Feria do Grelo” became “Feria del Clítoris”, which is ‘The Clitoris Fair’. The Galician grelo is a plant that is derived from the turnip. Hence, it has nothing to do with the female organ, but in Portuguese, the word “grelo” actually means ‘clitoris’.
Likewise, in 2018 a restaurant in Italy listed ‘lies’ on their menu due to an imprecise automatic translation. In fact, menta is the Italian subjunctive form for ‘lies’ but it also means ‘mint’, and it is evident the translation language programme used to create the English version of the menu couldn’t pick up the context in which menta appeared.
On a more frequent basis, this is particularly true when it comes to sayings that don’t have a perfect equivalent. Type ‘he’s the top dog’ on any translation website or app and select Spanish, you will get a literal el es el perro de punta.
Automated vs. Human Translation
So, will human translators be replaced by language translation machines? It’s evident that the simple, straightforward answer is no.
Check also this blog post about AI Voice Over vs Human Voices.
When should I use MT?
MT can be a useful tool for quick, daily translations, but it can’t replace human professional translators because it doesn’t pick up language hues as the human brain does.
And even when used for small, daily, unprofessional translations, it is advisable to have at least a minimal knowledge of the foreign language in order to spot inaccuracies.
FUN FACT! Did you know that in 2006 when Google Translate was released, the number of active users was in the thousands. Today it translates 100 billion words per day that is equivalent to 128,000 bibles worth for over 450 million people.
Have you enjoyed this blog? Check out a similar post on Google Translate’s Happy 10 Years. Or our previous blog, the most useful languages to learn for translators.
Remember, to discuss your next project, give us a call on +44 (0) 207 095 5730 or email projects@golocalise.com for a quote.