For many years, machine translation was associated mainly with laughable translation results. Not that long ago nobody dared to consider it a viable means of localising content meant for public consumption. But things have changed. The combination of consumers being used to getting what they want faster than ever before and the quickly progressing technology has enabled the machine translation market to truly flourish.
Services such as Amazon’s One Day Delivery have been born out of popular demand – the expectations of Amazon’s customers are that they can receive the product they purchase very quickly. Customers are less and less willing to wait for several days for their goods to arrive.
This attitude can be easily transposed to other services and aspects of everyday life. When you buy an expensive eye cream, you want to see the effects straight away without having to wait for weeks before you can spot the difference.
What modern consumers value is convenience and accessibility.
The translation world is heavily affected by this behaviour too. In an ideal world, there would always be enough time, money and resources to get translations done with the help of qualified linguists who are experts in their specific fields within a generous deadline. But, unfortunately, we do not live in an ideal world and anyone who has worked in the localisation industry knows that often you need to work around a number of different restrictions and still make the translation happen.
Machine translation benefits
Where timelines are short to the point of it being impossible for human linguists to turn around the required volumes, machine translation can be a real game changer. eBay uses unedited machine translation to localise their product listings on international sites. If we consider that eBay has close to 800 million listings of 300 words on average each and the listings constantly change – some are being removed, some are being added on a daily basis – it quickly becomes apparent that translating those with human resources would not only take years and years but would also prove very expensive.
What makes machine translation so attractive is that it can be done very quickly and cost effectively.
In situations where previously localisation simply would not happen at all due to time, budget or resource limitations, now thanks to machine translation, users around the world can gain almost immediate access to the content in their own language.
In that sense, machine translation continuously reduces language barriers and contributes to the knowledge economy.
Machine translation drawbacks
However machine translation is not the answer in every scenario. No matter how advanced a machine translation system is, it is unlikely to match the quality of a professional human translation. Therefore, if the translated text is of high importance and will have high visibility, it makes much more sense to opt for human translation rather than machine translation.
When it makes sense to use unedited machine translation
Although raw machine translation is sometimes used in healthcare communication, the military sector as well as in tourism and travel, unedited MT is above all a go-to solution for handling user-generated content. Many companies machine translate their online reviews, which have relatively short lifespan and translating them professionally would not bring much return on investment.
Other popular uses of machine translation include gisting and e-discovery. In simple terms, using MT for gisting purposes means machine translating a document from a foreign language into a language that is known to the reader to get a top level understanding of the source content.
Gisting can be useful whenever understanding the general message behind the text rather than understanding every word is sufficient. It is often applied to internal documents, RFPs, contracts and forms. Sometimes gisting can help to identify key parts of the source document that might be worth professionally translating if they need to be understood in full.
E-discovery refers to sifting through large repositories of data in search for the relevant information. Often e-discovery is part of fraud investigations, where several reports and financial data needs to to analysed to evaluate whether it holds any information that might be relevant to the investigation.
Although machine translation certainly has its imperfections and cannot always readily replace human translators, there are several use cases where it opens new doors to consumers and businesses, allowing them to gain quicker and wider access to information.