Machine translation is the hot topic in our industry. Especially since Google just lifted their self-imposed embargo on the Google Translate API.
Of course, this reignites fear and loathing among linguists and language service providers everywhere. I don’t share that fear. Why should we be afraid? Perhaps a loss of business? I can assure you I have no fear of losing the high-quality and high-risk work we do for our client that makes defibrillators or our client that makes heavy machinery or our client that makes rapid HIV testing kits or our client that makes four-story high printing presses or our client that makes bedside monitors or our client that runs a network of 100’s of hospitals and health care facilities or our client that….OK, you get my point.
Our industry is no different than any other industry. You have to match the right tool to the right job.
As of today, there is no machine translation tool, Google or otherwise, that will produce automatic translation suitable for use with customers.
However, there are many instances where machine translation is indeed the right tool. Here are some examples:
-internet chat with friends and family
-quick translation of a website by a customer who just wants to get the gist of a site (but let your customer decide to do that, don’t Google translate your site and pass it off as something you created)
-market research, marketers tying to do research on products in other countries can use MT for gisting, then when they want specific information they can have parts of what they found human translated
-low-risk research (patents, technical specifications, requests for proposals)
You will notice the only two-way form of communication I listed is internet chat with friends and family. No one else would forgive you when you say something so incredibly ridiculous in translation that if said to complete strangers would cause damage to any relationship.
In my opinion, MT should be used strictly for understanding and not for communication. You can certainly use the technology to get a high-level understanding of a document. Once that understanding is achieved you can decide if some or all of the documents should be translated by humans to get a deeper or exact understanding.