As a translator with a background in Linguistics—more specifically an MA in Computational Linguistics—I have great interest in the progress being made in machine translation. One cannot say that advancements in this field have been small or insignificant.
On the contrary, I’m really amazed that Google Translate sometimes renders perfectly correct sentences, under some circumstances. Years ago, this was something very rare, but it has become more and more frequent, depending on the material you are translating.
The type of machine translation carried out by Google relies heavily on huge amounts of information to train their system. As a result, it performs considerably better with general texts than in more specialized fields, where mistakes are noticeably more common.
You may be asking yourself: is Google Translate good enough for my business?
The answer isn’t simple. It depends on your line of work and the impression you want to make on your client.
Where Machine Translation Works
In some cases, Google Translate seems to be working well enough. A great example is the Chinese e-commerce giant Alibaba. They have literally millions of ads, and using human translators for all of them doesn’t seem feasible, especially considering how quickly their ads come and go.
Machine translation seems to be working well for them, particularly if we consider that (1) most product descriptions also have pictures, which will hopefully clarify at least some instances of mistranslation; and (2) consumers are often buying relatively cheap products. You wouldn’t negotiate the purchase of an expensive machine that costs hundreds of thousands of dollars based solely on a description generated by machine translation. However, the risks involved in buying a Bluetooth headset seem to be pretty acceptable.
… and Where It Doesn’t
But there’s another factor at play: the image that you want to convey to your client. When you buy something cheaply, you usually don’t get frustrated when you receive a product with a cheap package, or a poorly translated user manual. After all, that’s what you paid for. On the other hand, when you invest in a higher quality product, you tend to expect a better package and clear instructions.
The bottom line is that the translation you use has to match your product. If you sell cheap products, a higher rate of product defects is acceptable or expected. The same goes for flaws in their description, user manual, and so on.
If you sell medium- or high-priced items, you shouldn’t consider using machine translation, because it will produce errors that may upset your customer or harm your brand’s reputation.
Not sure you agree? Well, let’s look at some sentences translated using Google.
Google Translate at Work
From time to time, I like to try Google Translate and see how it’s progressing. Despite undeniable improvement, it often generates some curious or even funny sentences—well, not so funny if they end up on your packaging or as part of your product copy.
All of the following examples involve English to Portuguese translation.
These examples were all collected in the last two years. In other words, this is recent stuff.
Does Machine Translation Make Sense for You?
Ask yourself: are you willing to have these kind of errors associated with your brand? Translation costs money, but so do mistakes. Not only in terms of reputation, but often in the form of lawsuits from angry clients who feel deceived.
The point is that you need to weigh the pros and cons. Google Translate can be used when human translation isn’t feasible because of large volumes, as long as you clearly state that the text was machine translated. You also want to make sure that the amount of money you will save outweighs the potential harm caused by mistranslation and damage to your brand reputation.
Despite the great advancements in the field of machine translation, Google Translate certainly hasn’t reached the point where it can be safely used for high-end commercial purposes with no risk of blunders.
Matheus R. Chaud
I am a native Brazilian Portuguese speaker with extensive experience in translation, proofreading, editing, subtitling, and quality assurance.