Written by Mo Stone
As our world becomes ever more global and the bridges between people shrink, there is a greater need than ever for good, accurate translations. But just as technology has been the cause of this increase in need, it has also been the solution, as seen in the case of machine translation. Now that machine translation has become so ubiquitous in the industry, many wonder: could machine translation one day replace human translation? Read on to find out.
What is Machine Translation?
Machine translation (MT) is the use of artificial intelligence to instantly translate text from one language to another. Readers may be surprised to learn that this technology has actually been around for decades, but has only just come into widespread use. “I actually did my PhD thesis in MT and remember back when it was completely unusable,” says Eric Nichols, a senior researcher at the Honda Research Institute Japan. “My overall impression is that the quality has gotten much better than 15 years ago when I was working on it.”
We can see this improvement easily in the case of Google Translate. Who doesn’t remember the laughable nonsense of early translations using this product? But now, the technology has improved to such a degree that foreign language students and travelers rely on it, with decent results. From a business standpoint, MT is certainly enticing, and can reduce costs and increase productivity. Who doesn’t want low-cost, instant translations?
MT vs. Humans
With such leaps and bounds in MT, a common fear in the industry these days is that human translators won’t be needed in the near future. Says Nichols: “It’s always scary when a new technology comes along that looks like it’s going to take jobs away from people. And in some cases that does happen.”
However, there are still a lot of considerations with translation that machines just can’t do. Machines cannot recognize if a word is out of context, or if a phrase fits with a culture or not. Machines cannot capture style or tone that may be lost when translated from one language to the next. Human translators will always be needed, even if that means working in tandem with MT.
As Nichols explains: “I think in the case of translation, there are two roles: 1. writing the actual translations, and 2. doing the quality assurance to ensure the translations are correct. When people think of translation, they usually think of 1. and 2. gets ignored. And it may appear that machine translation can fill the role of a human translator in a lot more cases now, which might be true for producing a translation, but quality assurance requires a bilingual human, and as long as there are applications where a lot of money or risk are on the line, I don’t see that changing anytime soon.”
So while there will surely continue to be some cheapskate clients who choose to sacrifice quality in order to save on costs, as long as the quality of translations is important, human translators will continue to find employment within the translation industry.
How do you feel about machine translation? Is there a point we missed? Tell us in the comments below!