Do we hate on Google Translate?

People outside of the translation industry often mention Google translate to us with trepidation. Expecting us to have a genuine disdain for the service as it’s trying to “automate and take your business away” – In truth, this simply isn’t true, or at least not yet!

Yes, there may be the occasional smaller task that in the past may have been passed to a translation agency and now isn’t. However, Google translate is part of a modern digital tool set that gives ordinary people more confidence to engage internationally where previously they may not and this has to be seen as a good thing.

We at TranslationsInLondon use Google translate, yes you read it correctly! A translation agency using Google translate… is this not against the very principles we should abide by?! Well in short, no. We don’t have all possible languages in-house and if someone in our business development team picks up an email in a language they are not familiar with, what better way to quickly get the “gist” than to quickly run it though Google Translate? Is it a new potential client? Is it spam? Have we been left a fortune by a Nigerian prince that we need to quickly claim?! For these kind of quick ad-hoc information gathering translations, Google translate is ideal, it’s quick and best of all, it’s free. This is just one situation where a service such as this can be really helpful, there are surely many more.

I’m sure anyone that’s used Google translate for a something such as the above has then also had a chuckle or scratched their head with confusion at the sentences that Google translate doesn’t get quite right. The art of communication is an extremely complicated thing, the way we converse and construct sentences in the real world. We learn through years of practice to understand and adapt our language so more often then not the meaning or gist of what we are trying to say is obvious to another person who does speak the same language. For a computer algorithm to try and learn and implement these very human nuances is extremely challenging. A very straight up sentence with a literal meaning is a simpler task and the results are often very good. However, for a sentence or paragraph which is very descriptive and contains words with potentially multiple interpretations, the results can be… mixed.

Here is a nice quick culinary example sure to catch out some unsuspecting Italian restaurant owners:

Sott’oli misti della casa which literally means “mixed house pickles” – Google translate to this day translates this as “under the mists of the house” which sounds very much like something from a horror movie! Remove the apostrophe and you get “Mixtures of the house”  – better but still not quite what we are looking for!

This is not to say that the algorithms used are not being developed and are improving by the day. We have certainly seen a large spike in the accuracy over recent years. AI or artificial intelligence seems to be the next big technological movement and with this, companies such as Facebook and Google are currently investing heavily in technologies to try and improve the translation of status updates and news articles in order to increase the engagement of multilingual users and enable a singular piece of content to be available to the highest possible audience instantly. This technology is only going to improve year on year. Did we think 10 years ago that we would be on the cusp of driverless cars on our roads?

The trend however with these improving translation technologies seems to be very much focused on a familiar path. Improving the understanding of informative content for a potential consumer of this type of content. Throw away content is the wrong term but I think it emphasizes the point. For example, a breaking news article originally written in French that would be of great interest to a British reader. A largely accurate conversion of this text with the occasional incorrect syntax or noun would still allow for understanding and emotional engagement of the extended meaning. This is where machine translation really works. The original author of the article may not have quite the same opinion if the translation isn’t perfect but the point remains. 

Another thing that people SHOULD but often don’t know:

“When you upload, submit, store, send or receive content to or through our Services, you give Google (and those we work with) a worldwide license to use, host, store, reproduce, modify, create derivative works (such as those resulting from translations, adaptations or other changes we make so that your content works better with our Services), communicate, publish, publicly perform, publicly display and distribute such content” – So essentially Google can use your data as they wish! This may or may not be a huge issue depending on the specific document but if you have signed confidentiality or non disclosure agreements you may be in breach.

The real choice ultimately comes down to ingrained trust in the solution you are using. The written word is very much an art form and I’m personally not convinced a computerized system will ever be able to convey the deep meaning portrayed in many written works. As a general example, you’ve written your thesis, you have deliberated over the content for months are finally satisfied. You need to get it translated in order to achieve your dream of continuing your studies abroad. You could run it through Google translate and i’m sure the general concepts and ideas will mostly come across, but lets be honest you wouldn’t entrust this responsibility to an automated service because a) there is too much riding on it and and b) the tool isn’t THAT reliable to convey your message exactly how you intended. 

So ultimately Google translate gets a massive thumbs up from us, a very useful tool to be used in the right situations. Every tool has its place. Who knows what the future holds in terms of the world of machine learning and AI but we believe that the human touch will always have its place!

For more on machine translation, check out our blog post on MTPE

For gentle reading here here some Google Translate Fails!