The Truth About Translation Tools: What Works and What Doesn't

In 2023, the global language services and technology industry generated US$49.68 billion, according to CSA Research. That's a lot of money changing hands for words. But behind that figure lies a messy reality: translation tools are everywhere, yet most people misunderstand what they actually do. If you're a translator, a project manager, or a business owner trying to get documents translated, you need to know the difference between machine translation (MT) and computer-assisted translation (CAT) tools. And you need to stop believing that MT is about to make human translators obsolete.

What's the difference between machine translation and CAT tools?

Machine translation is fully automated. It takes text in one language and spits out text in another, with no human involved during the translation phase. CAT tools, on the other hand, are software that helps human translators work faster and more consistently. They don't translate for you; they assist you. The translator stays in control. Popular CAT tools include SDL Trados, memoQ, OmegaT, Smartcat, and Memsource. These tools use translation memory (TM), a database of previously translated segments that can be reused, ensuring consistency and reducing repeated work. So when you hear 'translation tool,' ask yourself: is it replacing the human, or helping them?

Can machine translation really replace human translators?

No. MT output often requires human post-editing, especially for idiomatic expressions, cultural nuance, and domain-specific terminology. Even the best neural machine translation (NMT) systems, like Google Translate and DeepL, make mistakes that a human would catch. The idea that MT will replace translators is a myth. What it does is change the translator's job: more post-editing, less translating from scratch. ISO 18587:2017 specifies requirements for full human post-editing of MT output and for post-editors' competences. So yes, you still need skilled humans.

Is BLEU score still a good way to evaluate MT quality?

BLEU (Bilingual Evaluation Understudy) was introduced by IBM researchers in 2002. It scores translations on a 0 to 1 scale by comparing n-gram matches between machine output and human reference translations. But BLEU is increasingly seen as unreliable because it correlates poorly with human judgment. Newer metrics like COMET and BLEURT are becoming standard. COMET, a neural framework developed by Unbabel, achieved top performance at the WMT 2019 and WMT 2020 metrics shared tasks. BLEURT, from Google Research, is a BERT-based metric that correlates better with human judgments than BLEU on several natural language generation tasks. So if you're still relying solely on BLEU, you're behind the times.

Do I need to use a CAT tool to be a professional translator?

Not necessarily, but it helps. CAT tools improve efficiency and consistency, especially for large projects with repetitive content. They also support standardized interchange formats like XLIFF, TBX, and TMX, which make it easier to collaborate with others. XLIFF is an XML-based bilingual document format standardized by OASIS; XLIFF 2.1 was approved as an OASIS Standard on 13 February 2018. If you're working with clients who use CAT tools, you'll need to be able to handle these formats. But for a one-off literary translation, a CAT tool might be overkill. Use the right tool for the job.

Are free online translators like Google Translate and DeepL good enough for professional work?

For gisting, yes. For professional work, no. DeepL Translator, launched on 28 August 2017 by DeepL GmbH, has reported that in blind tests with professional translators its translations were chosen roughly three times more often than those of Google, Microsoft, or Facebook. That's impressive, but it doesn't mean DeepL is perfect. It still makes errors, and for specialized domains like legal or medical, you need a human expert. Also, consider data privacy: free services may use your text to improve their models. For confidential documents, use a paid service or an on-premise CAT tool with MT integration.

What's the most important thing to remember about translation tools?

Translation tools are aids, not replacements. The best results come from a combination of human expertise and smart technology. Whether you're using a CAT tool with translation memory or post-editing machine translation, the human translator remains essential. So invest in the right tools, but invest more in the humans who use them.

Sources

• CSA Research - https://csa-research.com/l/media/Language-Services-and-Technology-Industry-Faces-Revenue-Decline-but-Remains-Poised-for-Transformation
• Machine translation (Wikipedia) - https://en.wikipedia.org/wiki/Machine_translation
• ISO 18587:2017 (ISO) - https://www.iso.org/standard/62970.html
• IBM Research - https://research.ibm.com/blog/bleu-nlp-benchmark-anniversary
• COMET (ACL Anthology) - https://aclanthology.org/2020.emnlp-main.213/
• DeepL Translator (Wikipedia) - https://en.wikipedia.org/wiki/DeepL_Translator