English Translations: Why Will Translators Survive?
According to Ethnologue, an authoritative guide, there are 7,159 living languages in the world. We at KLS Translation Agency translate into over 30 languages. However, English remains the most in-demand language for translations at KLS, in Ukraine, and globally.
We are fairly well aware of the reasons for this: English is the lingua franca of the contemporary world. It is the language of international cooperation and business communication, of the enormous scientific and literary content. English is the key language of Ukraine’s integration into the European Union, donor programmes, communication between arms manufacturers, and student exchanges. Everyone will agree that translations from English into Ukrainian and vice versa will only grow in volume. However, not everyone understands the input — both from the qualitative and quantitative point of view — of human translations to this body of work.
Many believe that the more important role the advanced AI plays, the more the work of professional translators shrinks. While people still recognise the role translators play in processing rare languages, they increasingly consider English as the almost exclusive domain of AI-based translations. Indeed, AI grants broad opportunities for communicating on everyday life topics, and many regard translations made by AI of low- and medium-complexity texts as satisfactory.
However, it is largely an oversimplification of the issue. And here’s why.
Machine translation systems are far from being perfect
Despite the improvement of machine translation systems, their shortcomings still persist, and they are far from always producing high-quality translations. The most popular Google Translate still struggles with long text fragments, idioms, or culturally attuned vocabularies. It can miss nuances, mistranslate gender, or provide word-by-word translations, losing the original sense.
More advanced machine translators, such as DeepL, are more accurate when it comes to certain languages. They can craft elegant sentences, sometimes obscuring errors inside, but of most concern are systematic biases, stereotypes, and cultural assumptions stemming from the biased data these systems are trained on.
Microsoft Teams and Google Meet introduce the option of simultaneous interpretation; however, it is far from being clear whether these tools have the ability to convey tones, nuances, or emotions, or whether they, on the contrary, reduce translation to just formal and superficial term equivalence.
AI mostly “thinks” in English
Another problem that is often disregarded is the dominance of English in the AI training models.
It is estimated that 90% or more of the training data that modern AI systems use is coming from the English-speaking world. This means that although AI systems have been designed to deal with many languages, they provide the best results and respond best in English. Of more importance is the fact that AI systems are used to English-speaking structured thinking. Put another way, they tend to process and generate languages based on the traditional English grammar, sentence structure and logic. This bias is not a purely technical one; it is a deeply embedded linguistic limitation affecting how AI processes all other languages.
Research shows that artificial intelligence models often demonstrate English-centred bias. They “think” in English, and only then provide translation in the target language, affecting the quality and nuances of AI responses in other languages.
Experts point out that the AI epoch splits users into those who just rely on it and those who proactively leverage technological advantages. When professionals review and improve translations of documents made by artificial intelligence, converting them into truly human texts, the results are much better than if people were to just use AI. It means that even in this field, human translators are irreplaceable in achieving high-quality and professional translation of content into and from English.
In the age of AI, it is translators who preserve the treasures of other languages
We should be aware that Ukrainian is not present among the languages that are actively used to train AI. It currently lacks the vast corpus required for AI training models. Its syntax is substantially different from that of English and other widely spoken languages. And its vocabulary reflects the cultural and spiritual heritage intrinsically linked to historical and contemporary contexts. That’s why people translating from English into Ukrainian and vice versa play a unique role in preserving the language, its diversity and richness, from “flat” machine translations. In societies like the Ukrainian, translators are the guardians of meaning, staying at the forefront of preserving cultural property and data sovereignty.
Hence, artificial intelligence does not replace translators, but modifies their role, making it more complicated. Today, they do not simply translate; they convey meaning, moods, exchange cultural senses, manage content, transmit irony, decipher wordplay and transform metaphors. They hear what has been said and what hasn’t. Machines are far from reaching excellence in all these realms.
That’s why the future of translation will not be machines replacing people, but machines complementing people. High-skilled professionals are switching from translating everything everywhere to choosing domains with high added value, like literature, marketing, and international relations, while retaining the leading role in technical and legal translations, as in all these areas, professionalism, creativity and ability to capture nuances are most valued.