Humans vs Machines: Which Translator is Better?
Language translation is an often necessary practice for various endeavors in the world. From negotiating foreign business deals to educational settings to understanding information contained in important documents, language translation is a trend that shows a potential of nothing but increase in the coming years.
Such a wide need for language translation has led one of the world’s leading technology juggernauts-Google-to experiment with and launch a language translation service. Debuting in the early twenty first century, Google Translate has become a convenient and thus widely used service to quickly translate various information into other languages, from single words to entire documents. Google Translate has recently added two impressive features to its translation offerings: visual and audio translation. The rollout of Google Translate’s new features prompted the translation service VerbaLink to compare the services offered by its human translators to the machine translator of Google Translate.
VerbaLink enlisted the talents of two of its employees for the test of the two translation services, professional translator Adriana to take on Google Translate head to head and quality control specialist Gaby to be the judge of the test’s results.
The study had Adriana and Google Translate work with two items to translate, both from Spanish to English. The first item was a written document about the benefits of beekeeping that was provided for Adriana and typed into Google Translate. Google Translate fared comparatively well with this portion of the challenge, accurately translating the lengthy and important words of the original Spanish. Gaby judged Google Translate to have adequately communicated the gist of the document. The service’s major shortcomings were in regard to proper English grammar and syntax. The machine struggled to translate the grammatical structures of Spanish to the grammatical structures of English, thereby producing more of a transliteration than faithful translation. The human translator Adriana, however, had no noticeable problems that Gaby pointed out. She was able to take the same Spanish text and convert it to a grammatically cogent set of sentences written in English. Round one went to humans.
The next challenge for Google Translate and Adriana came in the form of audio translation, a service in which the spoken word in one language can be translated to text in another. Google Translate accomplishes this task with voice recognition software while Adriana does so by hearing and then typing. Gaby was fairly impressed with Google Translate’s output but decided that it ultimately fell short in this round as well. One of Google Translate’s major flaws-as evinced by the graphic-was the machine’s misunderstanding of two words in the voice audio. These words were therefore translated in the English text as Google heard them in Spanish. Other weaknesses displayed by Google in this test were similar to those displayed in the first test. Namely, Google understood and translated the gist of the statement but fell short in regard to proper English sentence structure, syntax, and grammar mechanics. As a consequence, some of the information seeking to be conveyed was quite literally lost in translation. Adriana, on the other hand, produced a second translation that Gaby found flawless. She understood the spoken words and the context in which they were spoken. Round two also went to humans.
As this simple test suggests, humans should not have to relinquish the power of superior language translation to machines just yet. Humans are able to produce logical, grammatically correct sentences in other languages while machine translators still fall short in various respects. Still, the convenience of machine translators should not be disregarded, as they are useful for individual words as well as basic necessities when learning a foreign language, such as directions and locations. At the end of the day, humans remain the better resource for language translation.