Machine translation: where are we going?
In a recent US White House paper, A Strategy for American Innovation: Driving Towards Sustainable Growth and Quality Jobs, among the grand challenges one can find an “Automatic, highly accurate and real-time translation between the major languages of the world—greatly lowering the barriers to international commerce and collaboration”. This announcement has had a lot of buzz in media and in the industry. See also this.
Also some weeks ago, it was released that Facebook has filled a patent on their crowdsourcing translation platform and they are opening it for developers, which is another movement in this field since other companies such as Linkedin or Twitter have also proposed to be helped by (not professional) people instead of machine to translate their websites. Using (non-professional) crowdsourcing to reduce costs and avoid rates of professional translators is not the good option.
The debate is out there: man vs. machine. In our opinion, there should not be such a confrontation, machine help people, and people build machines to do things in a different and more advanced way. This sometimes hurts because people are reluctant to change. Besides, machines will not replace professional translator but rather help them earn more money! And this is the goal we as an industry have to face: increase the market size.
Thanks to recent advances such as translation memories and machine translation, where each translator once converted 2,500 words a day at a cost of some 25¢ per word, they can now offer 5,000 words a day at around 12¢-15¢ a word. However, the recession is making this industry suffer everywhere in the world. It is specially during these tough times where it becomes more important to reach the customer in their own language, i.e. the website, manuals, product documentation, and any press release, among others, should be localized. Localization increases sales!
Machine translation is widely deployed in big multinational corporations, including Microsoft, Dell, General Motors, Intel, or IBM, but the reach within translation agencies and freelance translators is still limited. The important thing is not to use generic tools such as Google (btw, it is great they add support to so many minority languages) but rather domain-specific solutions. Those companies are also spending a significant amount of money in professional translation services as well, and think also about the public spending in translation services. There is room for everyone.
With all this and this last article we share, we think that the translation market will boost if machine translation is widely deployed, the point is to keep the good ones and let the others leave …