Is Machine Translation ready for Financial Documents?

It is often tempting to rely on machine translation when it comes to financial documents, especially for large corporations and listed companies. Afterall, the documents and paperwork required to list and maintain a listed company are considerable. Simply upload your files onto the translation engine, and boom! Everything is done for you, without having to pay or even wait for the translation to come back from the agency, and it is only more of an irresistible idea during the reporting period, when speed has never been more important. But are machines ready to be of your service for financial translations?

It is said that language is an art which can only be mastered after thousands of hours of exposure and practice. Machines, despite capable of rendering simple and straight forward sentences, are too literal to handle texts that involve creativity. Although numbers constitute a large part in financial documents, there is often a need for companies to use figurative languages and sometimes with cultural context to build their corporate images. For example in financial reports and prospectuses, where corporations not only have to disclose their financial positions to the public, but also present information in a way to boost the confidence of their investors or potential investors, showing them that the outlook is promising. Can you imagine how badly a company can be affected if its slogan or tagline was wrongly translated? Take a noteworthy example of how the slogan of Hongkong and Shanghai Banking Corporation Limited (HSBC) “Assume Nothing” —meaning anything and everything is possible at their bank was translated into “Do Nothing” in many languages. In the end, the company had to spend an extra $10 million to clean up the mistake. Figurative languages can be difficult to human translators, let alone machines which can only rely on preloaded translations! Machine translations are not always wrong, in fact they can be useful sometimes, but are they powerful enough to handle creative texts that are far from literal? That’s the question.

People turn to machine translation with the belief that they can provide speedy translations, especially during reporting period, when there are a lot of documents to be translated within the same timeframe. It is true that you can find a quicker (almost instant) turnaround time on a machine translation, but you will likely have to spend valuable time reviewing and correcting errors. Translation engines might support a lot of languages, but according to research, the quality score of the translations generated by the most commonly used translation engines is only 30-40 in average, throwing doubt on their reliability. Consider a quick, yet perhaps inaccurate translation plus the additional time you will have to spend fixing it. Now compare that to the time for a certified human translator to translate your piece, check for errors and guarantee their final work. It might take slightly longer to get the final product, but the accuracy is worth it. Afterall, there should never be compromises on quality!