Machine Translation Blunders and Reality

  • 13 September 2016
  • by Szabolcs Kincse
Machine Translation Blunders and Reality

Machine Translation is a controversial issue for many in the translation industry. Let's see what you will have to say about the future of our industry: can man and machine live together in harmony?


I am sure you have heard the most embarrassing machine translation blunder of the summer: Microsoft was forced to publicly apologize after Bing’s own translation feature translated the word Daesh as Saudi Arabia instead of its real meaning ‘Islamic State’. To make matters worse, the news broke on social media soon creating havoc, anger and tension ultimately resulting in a Canossa of an extremely apologetic Saudi Microsoft boss.

It is commonplace that statistical machine translation can be effectively hijacked with little effort. Only a few thousand parallel translations are enough for the diligent algorithm to change the course of history and come up with a new translation.

If you have more and more translation work,
you might want to give memoQ cloud server a shot:
you need no downpayment,
pay monthly, and only when you need it.
Try it free for a month here!

The story however sparked some exciting new debates on professional translator forums about how useful Machine Translation is for the industry. Can the machine “learn to be human” enough to translate, say, literature? Will this technology be suitable to handle technical translations in the future?

While I do actually have a rather elaborated view on the matter (on what matter do I not?), first I decided to turn to you, dear learned colleagues. I would like you to answer a few questions, and to give your take on this exciting question. I will collect responses until the 26th of September, and hopefully, we will write an exciting post together on the matter. So please, respond, and provide whatever you think about the topic. In return I promise that I will deal with the subject duly and I will invite stakeholders, developers, users, companies to somehow show where we are at now, at the end of 2016.

Shoot away – here is the questionnaire: 

We have now closed this survey – thank you for your participation! We are now analyzing the results and creating a new post based on your input. Stay tuned!

The photo I used above is the work of Willie Williams taken at the rehearsal of Darren Hayes's Time Machine World Tour in the UK. (Own work, CC BY 3.0, Wikimedia Commons)