THE rise of translation engines or tools over recent years have offered users the convenience of getting a quick translation of a word of text, at the press of a button. And, the fact that they are mostly free adds to the growing popularity. The most popular amongst all has to be Google Translate, for the fact that the search engine is widely used on the World Wide Web today. As of May 2017, the artificial-intelligence- based multilingual machine translation service supports over 100 languages at various levels and serves over 500 million people daily.
But, an over reliance on translation engines and a lackadaisical attitude on the part of users, can put one in an embarrassing situation. We might have laughed at those hilarious images or signs that display poor translation, the conversion of words is so bad that you can’t even tell what the sign is saying in the first place. Poor translations can be funny in some instances, but they are not at all humorous when they put an entity in bad light or negatively impact a business image.
A case in point — a few weeks ago, a state agency was in hot water over an error-ridden notice in English that was put up at its office entrance. The notice was to inform the public that the centre would be closed for two days due to problems with the Internet and air-conditioners. The boo-boo prompted a ticking off from the chief minister who said any notice or announcement should be reviewed and scrutinised to prevent any mistakes. He added that text in English should not be produced through direct translation from any other language or by using translation engines available on the Internet.
There have been many other similar instances of machine translation fails across the globe, which have become viral on social media and the subject of ridicule by many. How do translation errors still take place, with so much technology at our fingertips? Shouldn’t it be a reliable tool to help with translation from one language into another? Apparently, not yet. One, it could be
that our technological advancement hasn’t yet conquered
language sufficiently to allow flawless translation that takes into account the meaning and context as well as actual word and phrase translation. Although major players, like Facebook and Google, are regularly announcing how far their translation mechanisms have come, reliably accurate machine translation is still some way off.
Machine translation, or computer-generated translation is, as is often reported, a technical marvel. It serves many important purposes, but it cannot yet replace professional human translators. Machines may understand how to directly translate a word, or a sentence word for word, but they are unable to understand the real context of a situation. The translation may not read naturally, contain plenty of grammatical errors, be poor in structure, and not convey the actual meaning. The reason that the accuracy and quality of machine translation is lower is because each language has a different structure. And, this factor cannot be recognised by machine translation apps — not yet.
So, with machines still a long way to go to produce accurate translation, it would be suicidal to totally depend on these engines to help you. It is, thus, good to observe the standard operating procedure when using translation engines. Always check and double check the translated text to avoid embarrassment and undue problems later. With the imperfections of man and machines, two “heads” are always better than one!
The writer is editor of BOTs, the weekly tech section in Life&Times. Trained in Maths, he has since traded his problem-solving skills with writing about how tech has helped to transform the world for the better