Yes, I want my consultation

Machine Translation: Chess Analogy

592436628_66fe64cb49_oLet’s talk about the future of translation. Not to make predictions, but to look objectively at some of the aspects where machine translation can have an upper hand. The famous chess game between the supercomputer Deep Blue and the then-reigning world champion Garry Kasparov provides an interesting perspective from which to do this. Deep Blue beat Kasparov in 2007, thus becoming the first machine to win a chess game against the reigning world champion. Can machine translation use strengths similar to those of Deep Blue to beat human competition?

Speed of thought

Just as Deep Blue could analyze 200 million positions per second, machines can retrieve translations from memory way more efficiently than human translators. Where the latter can spend countless minutes coming up with the right translation, a machine can find a translation in a matter of seconds. That does not mean, of course, that this translation will be correct or optimal. It will be as good as the database (often translation memories produced by humans) the machine has derived it from. But the sheer speed of retrieving translations puts machines light years ahead of human translators.

No human error

Deep Blue won the deciding last game after Kasparov had made a mistake in the opening. Translators are just as prone to human error as Kasparov was in that game. A translator who is tired may read “contact” where the original actually says “contract” or “from” instead of “form” (true story). A translator who is in a hurry may decide to forgo proper research and trust her intuition—and end up making a mistake. Or a translator may choose to guess the meaning of tags in her CAT tool segment instead of looking into the reference file—and arrange those tags incorrectly as a result. Obviously, machines are not prone to human error, giving them a huge advantage over humans.

Inefficiency is human

Whereas Deep Blue was 100% focused on the game, Kasparov likely lost at least some time unproductively due to understandable anxiety. In the same vein, machines are not prone to the inefficiencies and procrastination of a typical translator. Let’s face it: distractions and not knowing how to do things optimally often erode efficiency. When a translator spends countless seconds opening dialogs through menus instead of using keyboard shortcuts, he is wasting time. A distraction may steal your attention for several minutes or even hours. Whereas each loss of time appears small, over time, they build up to depressing numbers. This means less quality assurance and less research in the course of translation, all with a negative impact on quality.

In conclusion

When I come across a translation problem, I sometimes catch myself thinking that a machine would not have made the same mistake. Clearly, machine translation has serious advantages. It is our goal as translation professionals to stop perceiving MT as a competitor, at least for the time being, and incorporate it in the translation process to capitalize on its strengths. Read more about this in the article “Machine Translation Can Be a Good Servant”.


  • Shai says:

    Until true AI will be developed, the machines will do only as good as the humans who have programmed them and created the data which they use to emulate human actions. Therefore, the issues of human errors and “inefficiency” are all still there, although shifted to another stage of the work.

    I think that Chess is not a good analogy. Chess is a game with a very rigid rules the apply universally. It is easier to analyze potential scenarios within such controlled environment than in a less controlled environment. Translation is vastly different because there is no such thing as a correct translation. Literally there is (i.e. all the words are present) but there are other aspects in work here, such as the best tone considering the context, rephrasing an entire paragraph to convey the message, omitting and/or adding words or parts when it makes in the target language (and often it does for readability reasons), etc.

    Machine translation is a tool that can help professional translators in their work, and at other times cannot. It is not a human replacement. There are many who claim otherwise but this is simply for their own self-serving interests.

    • Hello Shai,
      Thank you for your comment. My goal with this post is not to say that machine translators can replace human ones, but simply to list some of the strengths MT has. I agree that today, machine translation is nothing more than just another tool in a translator’s tool kit.
      Best regards,

Add comment

About the Author

Roman Mironov
Roman Mironov
CEO & Founder

As the founder of Velior, Roman has had the privilege of being able to turn his passion for languages into a business. He has over 15 years of experience in the translation industry. Roman has helped dozens of clients increase sales by making their products appealing for speakers of other languages.