top of page

Super-Human Translators: The Finest Translators Are Also Great Writers

There are few aspects of life which have remained untouched by Artificial Intelligence (AI). Indeed, AI has become so sophisticated, and AI tools so widely available, that it is now hard to know whether the images we view or the videos we watch are real or AI generated.


The words we read may be AI generated too. Machine translation has been transformed by neural networks.

Could that mean that the services of human translators are no longer required?

What is the history of machine translation?

1954 was an important year in the history of translation as it saw the first public demonstration of what could be described as AI translation. In the Georgetown–IBM Experiment, 60 Russian sentences were translated into English using rule-based programming on an IBM 701 mainframe.


The experiment proved that machine translation was possible but was very limited in scope. There were no neural networks involved and so machine translation did not utilise AI as we know it today. The experiment was a technological advancement that certainly did not threaten the careers of human translators.


Across the succeeding decades, ever more advanced systems were evolved. In 1968 SYSTRAN was developed for the U.S. Air Force to translate Russian into English. It was the first large-scale, operational machine translation system.


By the end of the 1980s, IBM’s Candide Project had pioneered statistical machine translation (SMT), using probabilities based on bilingual text corpora rather than fixed grammar rules. It marked the shift to data-driven models. By the 1990s advances in global communications had significantly increased the demand for rapid translation and SMT became mainstream.


Initially based on SMT, Google translate was launched in 2006, bringing instant translation to all. Technological breakthroughs eventually led to the use of Neural Machine Translation (NMT). Google NMT was rolled out in 2016 and represented a huge leap forward in accuracy, fluency, and contextual awareness.


By 2020, multilingual AI and Large Language Models could handle dozens of languages with impressive accuracy and could translate idioms, cultural context, and even low-resource languages better than ever.


Has AI made human translators a thing of the past?

Some say it has but we are not so sure! AI systems are becoming increasingly sophisticated. However, even the most advanced of these still lack the unique qualities of the human brain and experience-based knowledge.

Here’s why human translators can and do deliver superior translations:


Writing skills

Behind every great translation is a talented writer. After all, a translation is, in essence, a rewrite of an original text. There’s a creative aspect to translation that AI cannot yet rival and perhaps never will. Good writers boast the innate skill to select not only the right words but those that are appropriate for the context of the piece. Words that conjure the required mood and inspire the readers.  Human translators effectively communicate ideas and concepts that can be lost in more literal translations.


Maintenance and adaption of tone

Most importantly, a high-quality translation retains or adapts the tone of the original material. Tone is a nuanced aspect of writing. Even sophisticated AI will struggle to discern the tone of a piece and replicate it in another language. In addition, a tone that is appropriate in one culture may not be suitable for use in another. Human translators can not only identify tone, they can also adapt it, if necessary, to suit the cultural demands of the new audience.


Contextual understanding

Humans understand subtle cultural references, idioms, and humour that don’t have direct equivalents in other languages. AI can seriously misjudge these aspects of language. Human translation will craft appropriate phrases for the target audience that prevent confusion or offence.


Nuanced decision-making

Translators make stylistic choices to convey the impact of the original, not just the literal meaning. They can decide when to keep a term in the original language, when to adapt it, and when to rewrite it for clarity.


Domain expertise

There is no substitute for human experience. Human translators are often specialists in certain domains such as law, medicine, literature or engineering. They lend their real-world experience to their work including their knowledge of specific terminology and how it is used in professional practice—not just what a dictionary says. They can ensure compliance with industry standards, something AI might miss.


Awareness of ambiguity

Many words and sentences have multiple possible interpretations. Humans can use context, author intent, and audience needs to choose the correct or most appropriate translation. AI often selects the statistically most likely option, which could be wrong.


Understanding of non-textual cues

Humans boast the ability to interpret meaning from layout, typography and tone. Non-textual clues influence how a text should be translated and cannot be accurately interpreted by AI.


Ethical and sensitive content

Humans can adapt language to avoid unintended offense or to meet cultural norms in sensitive contexts like healthcare, diplomacy, or marketing. AI might produce wording that’s technically correct but socially inappropriate.


In summary

AI translation excels regarding speed, consistency, and handling large volumes. Human translators excel in meaning, intent, creativity and cultural precision. By leveraging both machine translation and human skills, it is possible to deliver translations swiftly but without compromising quality. Computer Automated Translation (CAT) tools enable us to identify inconsistencies and eliminate human errors while human translators lend their knowledge, experience and creativity to projects.

Machine translation manages terminology and streamlines our processes. But science has yet to deliver the great writing that continues to make all the difference. 

Related Posts

  • White Facebook Icon
  • LinkedIn
Word Connection Logo

WORD CONNECTION SARL

60 chemin de l'aviation,

64200 Bassussarry,

France.

TEL: 00 33 (0)5 59 47 74 01

 

K.K. WORD CONNECTION
177 Benten-cho,
Shinjuku-ku,
Tokyo,
Japan.

TEL: 08 08 90 64 227

ATCC ISO 9001 & 17100 Certification Badge.png

Subscribe to Our Newsletter

Thanks for submitting!

Get In Touch Today

We'll be in touch soon!

Copyright © 2024 Word-Connection

bottom of page