Machine translation is playing an increasingly important role in business communications, according to one commentator. However, is it worth risking brand reputation on this less than accurate platform?

Karl Flinders writes in Computer Weekly that machine translation is now “critical” to many companies. It allows them to provide Information to colleagues, partners and customers all over the world in the fastest time possible. Social media has played a key part in this – with Facebook recently revealing it had acquired Mobile Technologies, the firm behind translation app Jibbigo – as it enables global real-time interaction. As a result, it is sites like Twitter and Facebook and their built-in translators that businesses are increasingly relying on to extend their reach. As a result, it is likely in this area where the greatest developments in machine translation will be made. As Mr Flinders notes: “Real-time translation of online conversations and comments will enable businesses to interact with billions of people, rather than millions.”

However, businesses might be wise to proceed with caution when it comes to relying too greatly on machine translators. The majority of the free versions, such as those used in social media and by search engines, are built on Statistical Machine Translation technology. This is a process whereby the search engine or social network scans the vast number of documents and correspondence it has at its disposal to analyse language matches. It will then identify the most common translations used between different language pairs and use this as the foundation of the translations it produces in response to user requests. While there is no denying the effectiveness of this method in generating results quickly, it is not known for its accuracy.

Anne Merritt, an English as a foreign language lecturer, wrote in the Telegraph this week that the major problem with Statistical Machine Translation is its struggle with syntax. This is what strings different words or phrases together to make a usable sentence. It means that while running a short Tweet or something written in basic language like a recipe through a machine translator will garner usable results, anything more complicated than that may not be reliable.

The problem is that if a business is using a machine translator to communicate with customers, they want to reflect themselves in the best light possible. However, if the machine translation they have used is – unbeknownst to them – gibberish, their customers may lose confidence in them. Indeed, sometimes automatic translations are no more useful than not having the message in the reader’s mother tongue in the first place. By communicating with clients in a less-than perfect version of their language, they might be left confused or even offended, and this could ultimately damage a business’s brand.

Yet this should not put off businesses from investing in language services. Numerous business commentators attest to the importance of communicating with a target market in their native language and it is known that customers prefer to buy products and services when purchasing in their mother tongue.

Of course, businesses could just focus on the territory where they are based, but this would be missing out on a huge opportunity. Thanks to the internet and digital communication, it has never been easier to work with manufacturers on one side of the world and attract customers living on the other. As Mr Flinders explains, globalisation means businesses today “have to communicate internally and externally in different languages”.

However, having access to translation services is not only an important part of attracting new customers, but also communicating with current ones. Mr Flinders says companies need to be able to get involved in conversations about them on social media, and if they cannot follow what consumers are saying they will not be able to. As a result, important opportunities may pass them by.

Machine translation can be employed in these instances to get the gist of what is going on in social media in a variety of languages. It means that marketing teams and customer services can get involved in the conversation – which goes a huge way in cementing client loyalty, particularly when advice and assistance is being offered in the person’s native language. Yet at the moment, for anything more complicated than this businesses would be best advised to employ a human translator.

The developers behind machine translation technology continue to make breakthroughs, and there appear to be two major areas of focus for the future. One is getting to grips with translating syntax, so that the technology is able to make sense of longer pieces of text. The other is improving the machine translator’s capabilities in more languages. There is an abundance of English and Spanish language text online for Statistical Machine Translators to work through and learn, which is why translations into or out of these languages are often the most accurate. Meanwhile, other languages like Mandarin are less well covered online, so the accuracy of translations in these languages suffers.

Until these improvements are made, human linguists should continue to be a business’s first port of call when a translation is required. However, even once the technology is developed further, there will always be a place for professional translators. It takes a human to correctly localise a translation so that it is not only fluent for the target reader, but also compelling and able to prompt the reaction that was intended. This means hiring a person who lives in the country where the target language is spoken – so they are familiar with the latest terms and words – and who has experience working in your industry. It ensures that any technical terminology will be translated correctly.

Real-time translation technology is obviously already playing an important role in business relations and will continue to do so. However, until it has reached a level where the results are verifiably reliable, technology should never be a company’s sole consideration.