Never let a machine do the job of a human...

This week, a so-called breakthrough in real-time, automated translation has made the headlines in the languages and translation sections of the international press. Not in a particularly flattering way, though.

The new technology is called “Lexifone”, and it’s main innovative feature is that it is not a machine translation along the lines of Google Translate or Babelfish per se, but rather a real-time interpreting service for eight languages that works with “nothing more than a phone”, according to information provided by the Associated Press: “The service translates spoken conversations in real time, which Lexifone says is an improvement over free, Web-based services that are typically limited to typing in text”.

Practically all Science fiction shows and movies that involve space exploration feature a highly advanced, real-time translator devices that allow seamless communication between entities from different planets –not to mention countries-, and we have now reached a point where we don’t even question the complexity behind such a technology. But when it comes to the present and our current level of technological prowess, an announcement like the one of “Lexifone” does sound quite far-fetched. Mainly, because it involves two fields where we, translators and transcriptors, see how technology fails miserably every day: machine translation AND automated speech transcription. The first, albeit useful for separate words and simple sentences, proves useless when it comes to idioms and intricate text structures. As for the second, just try throwing in a strong accent – then sit back and have a laugh watching the machine spell utter nonsense.

Those two factors already make me feel somewhat skeptical about this particular software’s functionality. But what really makes me dismiss it as a potential tool for my job as a translator is the fact that it emphasizes its low costs: “Lexifone allows people to get translations without paying hundreds of dollars for human interpreters”, writes Associated Press. Specifically, the service costs 15 to 40 cents a minute.

Now, that is plain suspicious. I know that we all live in a world where we want more for less, but one thing still stands: if, apart from quantity, you want quality in your translations, it comes at a price. And even if we are talking about a machine, there should be people behind it constantly working on improving its interface, its voice recognition system and its vocabulary banks, trying to keep the languages as fresh as possible every day. And that should cost quite a bit of money.

Unsurprisingly, the Associated Press goes on in its article to say that “tests conducted by AP reporters in Mexico, France, Israel, China and the U.S. show that Lexifone is still far from delivering the quick and seamless translations it advertises. Using the service proved frustrating, both in the quality of translations and the length of time it took to complete phone calls”. It shows promise, AP writes, and it could help in emergencies, but it’s definitely not fit for everyday use. And whereas the project deserves praise for the idea and the effort, the truth is, a machine will never beat the quality and flexibility of a translation job done by humans.

Because, just as humans, languages are alive. And in inWhatLanguage, we acknowledge it.