Can machines outperform humans at translating?

Ian Abernathy
Even with major advancements in technology, human translators are still a force to be reckoned with. In the battle between human and machine, who comes out on top? See why the future of translation may be going in a different direction than you thought.

In recent years, the increasing automation of services has threatened to replace the skilled handiwork of writers and translators. But in the same way that early films incorrectly predicted intricacies of the near-future, technology has yet to outperform humans in many fields that were surmised to be overtaken. American author and essayist Paul Goodman summed up the necessity for human translation long before Google Translate was a lurking inevitability. He wrote, “To translate, one must have a style of his own, for otherwise the translation will have no rhythm or nuance, which come from the process of artistically thinking through and molding the sentences; they cannot be reconstituted by piecemeal imitation…”. Herein lies the secret to translation: there is no replacement for the intrinsic voice of the human translator. No matter how adept a machine may be at rendering words into different languages, context is often thrown to the wayside.


Despite the proven efficacy of human translators, machine translation systems are still being developed at a rapid pace, going up to bat against human translators in competitions of skill and speed. Several empirical studies have been performed, matching up translation services like Google Translate, Systran, and Papago with human contenders. Both human and machine are then tasked with translating a series of texts into different languages.
The results of one particular experiment found the three aforementioned tools to each score less than half the total points possible for a “successful translation.” Humans did much better – averaging a score above 80%. When combing through similar experiments online, all data reveal a consistent trend; while machine translators complete tasks at a rapid pace, they fail to retain the nuances, flow, and tone of each source text in the new language. Human translators, while much less efficient, are more adept at treating texts as a whole, contextual pieces, not simply a string of words devoid of connotation.
Fail to retain the nuances, flow, and tone of each source text in the new language.


Automatic translation isn’t wholly ineffective. Indeed, there are some instances where such applications can clear up a misunderstanding, such as an email sent from a foreign party. In essence, automatic translations are ideal for shorter texts that are not meant to be published or distributed. Systran, which boasts itself as a “pioneer” in translation services, claims to help organizations better their productivity concerning multilingual interactions. For businesses looking to communicate with potential clients, this could be an effective, real-time tool.
However, the usefulness of this software is not to be overestimated. Text often still sounds awkward or unclear. Machine translation can still endanger a professional global image for your brand. Even Google does not support using machine translations for multilingual content marketing. In fact, the juggernaut will punish this kind of behavior. Companies looking to boost their brand presence using SEO tactics should avoid machine translations at all costs. Failure to do so will likely make your website look amateur, diminishing the potential of increasing your conversion rate.


For businesses hoping to expand into global markets, translation is a very real issue that requires the utmost care. Beyond this is the importance of localization – translating content from the source language into a new, culturally-specific dialect. Localization involves translation, but it is much more than that. In order for content to be truly localized, it must be optimized to a local language from the perspective of a culturally-sensitive and informed speaker. Concerning machine translation, there is simply no tool, at least at present, that contains both the innate and learned fluency of a local language. This is because language is more than just diction itself. Words carry culture within them – something Google Translate just doesn’t get.
As part of a multilingual content marketing strategy, localization is the word on every successful company’s lips. Localization encompasses design and color, typeface and video. In fact, it includes anything on an app or website that you wish to tailor to your target audience. Even subtitles and voiceover require proper localization to effectively portray a message. This further reveals how translation is a layered, complex approach that is only championed by human experience. Furthermore, global customers will only take seriously those products and services that are sufficiently adapted to their location.
Localization encompasses design and color, typeface and video.


If machine translations trailed human skill by a mere margin, we might suggest translators start searching for new work. As of now, though, we see a clear disparity in quality between the complex translations of humans and the “piecemeal” rendering of the machine. With that said, it becomes increasingly important for businesses to employ the services of native human translators. Especially with localization, native translators retain the most well-rounded understanding of a local language. As the online market becomes more saturated and SEO techniques remain ever-important, effective translations are integral to any modern, global content marketing strategy.