Videofix: Why a debate on machine and human translation at all?

March 29, 2017 3:32 pm

The debate over the idea that machine translation will replace human translation has been running for a long time, but is there a reason to debate at all? (Human) translators seem to share a common stand: humans will always be needed in the translation process, at least to proofread the results of machine translation. But there are other voices coming from the IT sector that claim the opposite.

Videofix_ Why a debate on machine and human translation at all_

It is the case of Ray Kurzweil, expert in Artificial Intelligence, who explains that computers will soon achieve human levels of language understanding that will allow them to deliver top quality translations, at the level of the best-possible human translation:

Hilarious mistranslations performed by statistical translation software circulate among professionals in an attempt to prove that this is never going to happen. However, if we take into account the recent developments in machine translation and CAT tools, e.g. at the EU institutions, we can at least confirm that the role of translators will change a lot in the long (or maybe not that long) run.

One of the key issues raised when engaging in the human vs. machine translation debate is to what extent machine translation quality will meet human translation quality, i.e. if we will really manage to make computers achieve human levels of language understanding and if so, will this mean the replacement of human translators as we know them today. However, this argument disregards the fact that translation quality is highly-subjective concept and that translation is not only about bilingual equivalence, but also about context, meaning-conveyance and language for specific purposes or specialised language. Languages are obviously required in translation but they are one instrument within the language-mediation toolbox, and this is why bilingual individuals do not necessarily make the best translators. So should we really care to debate over machine and human translation at all? Only time will tell.


Written by Doris Fernandes del Pozo – Journalist, Translator-Interpreter and Communication Trainee at the Terminology Coordination Unit of the European Parliament. She is pursuing a PhD as part of the Communication and Contemporary Information Programme of the University of Santiago de Compostela (Spain).

Sources:

  • Davis, Adam (2014) “22 Hilariously Inappropriate Mistranslations”, BuzzFeed. Available at: http://bzfd.it/2o6JKRO (Accessed 28th March, 2017)
  • Kelly, Nataly (2011) “Ray Kurzweil on Translation Technology”, The Huffington Post- The Blog. Available at: http://huff.to/2neEc4i (Accessed 28th March, 2017)
  • Kelly, Nataly (2014) “Machines Alone Cannot Solve the World’s Translation Problem”, The Huffington Post- The Blog. Available at: http://huff.to/2njb3Vi (Accessed 28th March, 2017)
  • McDonald, Fionna (2015) “The greatest mistranslations ever”, BBC Culture. Available at: http://bbc.in/2ogvUtE (Accessed 28th March, 2017)

2,308 total views, 6 views today

Categorised in: , , ,

  • David Fernandes Del Pozo

    I think translators do not have to worry too much on this issue. The reason is that humans need to adapt depending on the current machine skills to translate properly. Hence, human translators should not be afraid of this change but really promote it. Not because you can perform calculations by hand you are suddenly throwing away your calculator.

    • termcoord

      Very interesting contribution and example indeed, David. Thank you for reading us!

  • Walid Saba

    This is what happens when pseudo scientists speak about one of the most difficult problems in cognition/artificial intelligence, namely language understanding. Statistical approaches (as well as all data-driven approaches, such as deep neural networks, etc.) are doomed. Language cannot be learned from observing patterns in data. Moreover, before they do translation, let them show that they understand simple reference resolutions that a 5-year that never processed/heard millions of terabytes of text can effortlessly do. Little knowledge is truly dangerous. This misguided hype must stop, sometime

    • termcoord

      Thank you for making such a sharp observation, Walid Saba. This is why we ask ourselves if we really need a debate over machine/human translation. Thank you for reading us!