A Translator and the IT Translation Revolution

My name is Marco, and I’ve been a translator for almost 20 years. When I was younger, I tried translating books that weren’t available in Croatian, and gradually, translating became not only my hobby but also my primary source of income.

As a student, I freelanced, taking on individual projects at translation agencies, and eventually started working as a translator in various companies, both international and domestic. I often encountered new and interesting tasks — from translating technical documentation for international partners to adapting marketing materials for foreign markets. One of the most memorable experiences was live-translating a videoconference with about 30 participants, collaborating with several other linguists.

Over time, though, the work started to feel repetitive and routine. Translating similar types of texts wasn’t too challenging, but it felt draining, as if the inspiration was fading along with any sense that my work was making a positive difference in the world.

At the same time, I used to look down on machine translators — to me, Google Translate was synonymous with sloppy, clunky translations. I still believe that nothing can replace a well-trained human translator who understands context and language nuances, especially for complex tasks.

A Shift in Perspective

I remained highly critical of machine translation technology until about three years ago, when a colleague sent me a machine-translated text, apparently about robotics, with a comment: “Look, soon these metal monsters will put us out of work.”

The text was almost flawless, and I started reading up on what had changed in machine translation to improve its quality so much. It turned out that, thanks to the development of neural translation over the past decade, modern machine translators like Lingvanex, DeepL, or Microsoft Translator can now handle most tasks with about 80% accuracy. For many companies, this level of accuracy is enough; they don’t always need a perfect, ultra-precise translation. What matters more is that machine translation can save a lot of time and resources.

Modern programs analyze entire phrases rather than individual words, trying to capture context and semantic connections. The result is a natural and smooth translation.

However, there’s still a long way to go before perfection:

1. Machine translation still occasionally misses context, which lowers the quality of the text or sometimes severely distorts its meaning.

2. Machine translation doesn’t always consider cultural nuances (like etiquette norms).

3. It’s prone to errors in conveying figurative language and style.

These limitations make it essential for human specialists to review even high-quality machine translations.

Now, I regularly use machine translation in my work but meticulously review the output. The primary responsibility for translation quality still rests with the human translator, while technology significantly lightens the load and speeds up the process, freeing up time for creative work.

In short, my view of machine translation has changed dramatically, and using it has opened up new opportunities.

Essentially, new technologies don’t take work away from highly skilled translators; they free them from routine tasks, giving them more time for intellectual work.