Google Translate is still making the news on an almost daily basis. Sometimes it’s for the major technological advances it has made and its impressive new language translation capabilities. Other times it’s for the major mishaps it’s linked to and the fact that it often causes more problems than it solves.
So how much has Google Translate really come along, in terms of its machine translation abilities, and to what extent is it still lacking? Let’s take a look.
It’s not all bad…
There’s no doubt that Google is making strides in the machine translation field. Its camera technology is astounding, for example. Not only can users now point their phone camera at printed material in a different language and have it immediately translated in front of their eyes, but Google Translate now automatically identifies the language for the user.
This means that you don’t even need to know what language you are translating, and the whole process is a lot quicker as well. Just what you need when you’re desperately trying to understand railway signage in Kuala Lumpur, as your train is due to leave the station.
Its computer vision translation tool is perhaps Google Translate’s most celebrated technology and an additional 60 languages were added to its arsenal earlier this year. As a result of this significant boost, there are now very few instances where users would find themselves without access to the tool in their desired language, and that’s pretty impressive.
However, what’s less impressive is that its accuracy is still just not up to scratch.
The Wikipedia debacle
Wikipedia is, unarguably, a pretty amazing online tool. For English speakers anyway. We have over 5.5 million articles to peruse and search. However, the same cannot be said for other ‘editions’. In fact, only 15 of the 301 editions that are in other languages have over a million articles.
Understandably, Wikipedia has recently realised that something needs to be done about this and brought Google onboard to help with translations. Google Translate was incorporated into the Wikimedia Foundation’s existing content translation tool and it was intended to improve it with the best machine translation available.
However, critics (including Wikipedia editors) claim it has caused more problems than it’s solved. Editors working on non-English editions started using the tool to translate articles from other editions. They began to see major problems almost right away.
The value of Wikipedia is its accuracy and strange and obvious errors popping up mid-article can damage its reputation fast. In order to try to limit the damage, Wikipedia launched a tailored deletion tool to help editors cut out poor translation quickly and easily.
In conclusion, although machine translation can compete with human translation in some specific circumstances, these successes are usually for very narrow examples that require little true understanding for linguistic nuance.
Businesses embarking on major translation projects still benefit from working with human language translation service providers who can provide so much more than like for like, literal translations.
While Google Translate can be a life-saver in that Kuala Lumpur railway station, it’s still very unlikely to cut the mustard on your next major international marketing campaign.
If your business is looking at branching out into new overseas markets, you need to speak your new customers’ language. Talk to the language translation experts at Bubbles today and find out how we can help.