Companies and machine translation: human translation still better?

A recent article in The Wire by Ian Henderson dealt with the naive view of applying machine translation tools rather than professional translation companies and translation services. Whilst machine translation can be a time saver for quick understanding and useful when you’ve lost your way abroad, the technology still is far from accurate for fast businesses. According to Henderson, chairman and CTO Rubric, a lot has happened in the machine translation industry. It has now become a buzz word. He mentions a few new developments we have discussed in our blog, such as the simultaneous voice translation service Microsoft invented for Skype and Twitter’s new translation tool with which soccer fans from all over the world were able to tweet to one another during the World Cup in Brazil this summer. He also elaborates on Google’s acquisition of an app called World Lens which enables users to translate signs or menus in foreign languages. Henderson argues that World Lens merely gives an estimation of what a sign or menu says, not what it actually says. He has tried out the app himself and came to the conclusion that it was what he expected: a great app for tourists that are struggling to get around in a foreign country, but it is definitely not suitable for businesses. After all, who would want their product to be accompanied by the Spanish-to-English translation “Take to end the following procedure?” Are Computers Becoming Translators? Henderson believes Google Translate has caused automatic translation to become a mainstream phenomenon. Moreover, as huge developments are made in the technology in general, many people believe it won’t be long until an automatic translation service is created that can tear down all language barriers. Henderson disagrees: languages are more than just words, he says, which means translation machines not only have to do mathematical calculations, but should be able to carry out abstract thinking as well. Thus, he says, businesses must think twice before they decide to use automatic translation when entering the global market. There is a difference between consumers’ expectations of free tools and products they have to pay for, Henderson states. He compares translation technology to Youtube: this video service started out as a way for people to show their homemade videos to the world, but over the years, the website has seen a transition to more professional videos. Henderson believes apps such as World Lens are also right in between homemade and professional. It might make choosing a restaurant in a foreign city a little easier, but it won’t be of any use when you are trying to assemble manufacturing equipment. Similarily, Google and Bing Translate take the same position in the translation field: according to Henderson, this became very clear during the live demonstration of Skype’s real time translation tool. When Skype’s vice president Gurdeep Pall called one of his managers in Germany to demonstrate the service, their conversation was hindered by delays in the translation, he says. Henderson believes the confusions that occur because of these delays don’t matter in casual conversations, but are harder to overcome in business meetings. The Positives of Machine Translation Even though it has its downsides, Henderson thinks machine translation can be of great help to human translators. Many global language service providers use machine translation to create translation memories that can be used by humans to speed up the translation process. This creates a situation which has the best of both worlds: the human translator brings his cultural knowledge to the translation, while the automation technology ensures the task is carried out quickly. The human translator’s cultural knowledge is exactly what is missing from automatic translation services, Henderson states. A casual graphical theme, for example, can work great the US, but will probably fail to appeal to a German audience, which expects a something a little more formal. Thus, Henderson concludes, automation technologies might be of great help for consumers that are looking for a quick translation, businesses should look for other options. Marketing strategies differ per audience, and the nuances accompany this cannot be translated by using translation technologies. As long as automatic translation tools cannot spot cultural elements in a text, Henderson says, translation machines won’t be able to translate language completely.

Carla Estefanía

11/30/01