Man vs. machine

Google can now translate almost anything. But for many things, the translator's better than the algorithm – and that's unlikely to change anytime soon.

Our Head of Operations in Zurich, Patricia Kamer (picture), recently published an article in the “Werbewoche”. We wanted to make it available for you here on our Website:

By now, most people will have used Google Translate, DeepL or some other machine translation tool to quickly find out what lies behind those interesting-looking Chinese or Russian characters, or to understand what their French colleagues are saying in that email. The speed of the tool is what really catches the eye – that, and the fact that the services are often free.

The counterpart to machine translation is transcreation: the term has its origins in the advertising sector and once described the translation of advertising claims. In addition to the ability to translate, the ability to create has always called for a great deal of inventiveness – it's where the translator shakes off the shackles of the source text. The idea is to create a message that triggers the same emotions in the target audience's culture as those of the original. Nowadays, the term is sometimes used for translations that flow as naturally as possible.

Translators, agencies and language service providers are finding themselves between these two concepts – transcreation and machine translation – and having to ask themselves whether the groundbreaking improvements in machine translation are destined to make human translators redundant. Today's neural machine translation technology leverages artificial intelligence: these "adaptive" machine translation tools learn from humans how to translate correctly. The results of the aforementioned DeepL translation tool can give even the most critical linguist pause for thought.

Are translators destined for the scrap heap?

The question is whether the new technology works properly. So, what situations call for human capabilities? That depends on what you want to achieve – the purpose of the translation will determine which path you go down. Basically, if all you want to know is what a text in a foreign language is saying, machine translation is a handy tool. But if the text contains complex formulations, or you need a result that's stylistically correct and doesn't sound like a translation, then it's advisable to use a professional translator from a translation agency. The same goes for texts that are confidential or contain legal or etymological concepts: use native-speaking professional translators specialising in the relevant field. And in terms of data security, you're better off having your texts translated by a translation agency than doing it yourself online.

Machine translation and transcreation

If it's a question of simply rendering a text into another language, the coming years will see machine translation take over from human translators. But machine translation has its limits, especially where a text needs re-jigging for the target audience, or where it contains local colouring or special terminology, or where the author wants it to have a particular effect on the reader. One day, maybe, we'll no longer speak of "translating", but of machine translation versus transcreation. Then, at least, people will know what to expect.


Yes, I am interested in receiving news about Apostroph. Please send me your newsletter.

Newsletter archive