HUMAN VS MACHINE TRANSLATION: WHICH ONE YOU SHOULD CHOOSE
by Diego Cresceri
Technology has reached unprecedented levels, and progress is not doomed to slowing down. Machine learning and Artificial Intelligence promise to disrupt the way we live in different areas, and the translation industry is affected as well.
Tech progress brings about many opportunities to companies that need to translate large amounts of content in a short timeframe. As we said in this post, machine translation allows for extraordinary results, but human translation is not replaceable (at least not in the medium term), especially for some content types.
This all being said, when it's a good idea to choose Machine Translation and when human translation is to be preferred instead? Here’s what we think.
WHY MACHINE TRANSLATION?
Today more than ever, global companies compete on quantity (as well as on quality) of content. Inbound marketing strategies require creating and localizing a vast amount of content to be published on the Web in order to attract traffic to their website.
And there is more: if you want to tackle new markets, localizing your products or services into the languages spoken in your target markets - including all collaterals and support documentation (website, installation and user manuals, post-sales service documentation, and so on) – is paramount.
This exponential growth of content necessarily requires the implementation of less expensive localization strategies which, at the same time, ensure more acceptable timelines. Machine translation can meet both budget and timing targets, which leads us directly to the next topic.
Need for faster time-to-market
As we said, technology evolves relentlessly, and time-to-market of new products or solutions can be an important advantage to keep one step ahead of the competition.
A serious machine translation program, which includes mindful authoring (for example by using the so-called controlled language), pre-editing and post-editing by native speaker professionals, allows a much faster time-to-market compared to traditional translation programs.
Fast-growing new kinds of content
Corporate blogs, patents and scientific publications, user-generated content, customer support and customer care: these are just some of the new kinds of content that have been growing at an extremely high pace in recent years.
More and more marketplaces, most notably eBay, now offer global users the possibility to view all ads in their own language. Similarly, many travel and hotel booking providers automatically translate guest reviews into all languages.
This is an important service, which allows for more informed purchase decisions. However, its cost would be prohibitive without automated translation.
ADVANTAGES OF MACHINE TRANSLATION
Where can machine translation actually beat traditional, human translation?
1. Terminological consistency
If properly trained, automated translation software can ensure terminological consistency to a greater extent than human translators, thus enabling a consistent user experience. For this to happen, however, an initial investment on engine configuration is required so that it "learns" the relevant terminology.
We all have a word that we just cannot type correctly on the first try. I often misspell words ending with "-zione" when I'm writing too fast. Automated translation programs, by retrieving phrases and words from a database, cannot make this kind of error (provided there is no typo in the database, but that would happen because a translator had left a typo in a sentence that was later fed into the database). Machine beats human on this one, no doubt.
We have been mentioned this already: based on industry standards, a professional translator can translate an average of up to 2,500 words a day, even less in case of specialized texts (some translators have a productivity of 10,000 words per day, but let’s keep the feet on the ground). Machine translation engines can translate thousands of words in a few seconds and productivity for post-editors is well above translation (with peaks of up to 1,000 words per hour). For contents where post-editing is not foreseen (this is the case with some large IT companies’ service portals), the advantage in terms of timing is overwhelming.
As we have seen, a serious pre-editing + machine translation + post-editing strategy can lead to considerable cost-savings. Machine translation allows localization of content that would otherwise be left untranslated due to cost constraints, and can represent an important competitive advantage.
WHEN SHOULD YOU CHOOSE HUMAN TRANSLATION?
Once again, human translators are not heading towards extinction. There are at least two reasons why even the smartest machine will never replace a good human translator.
Some kind of texts, by their nature - i.e. newsletters, landing pages, and marketing content in general - have to be creative. In order to be compelling, for instance, marketing content cannot rely on controlled language, which is one of the prerequisites for effective machine translation. No existing software can beat a good translator on creativity.
2. Knowledge of target culture
Machine translation is not able to recognize target culture, nor to interpret and localize any word pun. When source content contains cultural references, puns, and anything that requires true localization, or perfect knowledge of the target culture (for example in order to avoid offensive, irrelevant or incomprehensible element), the work of a native speaker is essential.
It is not just a matter of numbers, then: not all content is suitable for machine translation, with or without post-editing.
If you do not know which strategy is best for you, contact us and we will help you to assess your content and propose the best solution.
Diego Cresceri - Founder and CEO of Creative Words, he does not deny his past but never looks back. An absolute lover of languages, he's an incurable optimist and cannot wait to se what the future holds.
"A different language is a different vision of life." Federico Fellini