In the last few months, the matter of artificial intelligence was thoroughly discussed and scrutinised as deeply as possible. However, it still stirs great controversy and inspires fear of a new technological revolution which is happening right before our eyes and might have consequences that are hard to predict.
The tool which initially was meant to help us with repetitive tasks turned out to be surprisingly good at tasks which require creative skills. Thus, it hit sectors where it was least expected – and among others, translation and linguistic services. Do chat GPT and other chatbots really pose a threat to the whole lingustic sector?
Machine translation and artificial intelligence
Although artificial intelligence (AI) is often confused with machine translation, in fact the matters are much more complicated. The term “machine translation” (MT) might evoke legendary translation fails generated by Google Translate in its early days – which is correct, but these times are long gone.
One of the first machine translation systems used statistical data to generate translations (statistical machine translation – SMT). The software compared a range of translations and used them to create a new translation in a given target language – word by word and phrase by phrase. However, this method has many disadvantages: it doesn’t recognise the context, the tone of the message or subtle linguistic nuances. Hence, it can’t be applied to texts of specialist, scientific or literary nature, which are usually highly complex.
Another option is neural machine translation (NMT) which draws from a completely different technology of creating so-called neural networks resembling the scheme of human brain to deliver a translation. This technology is used by new MT software and in the AI systems. The tool analyses the whole text and recognises the rules of the given language, context and style. It’s able to convey nuances, humour and irony. In short, NMT is a tool that has its own rules and can be developed in many ways thanks to the deep learning technology. What’s more, the software learns at an exponential speed, which makes it impossible for humans to compete with it. That’s why AI rocks the foundations of the labour market – not only does it allow for cutting costs of employment, but it’s also way more efficient than an employee (or even a whole team), eliminates typical human errors and faults.
Chat GPT, Google translate, DeepL or… a human translator?
This is not the first technological revolution which strikes the translation sector. We’ve already gone through the same controversy and concern when CAT tools became popular. However, computer-assisted translation didn’t render translators obsolete. On the contrary: today, it’s a priceless support in their daily work, which guarantees terminological consistence and high quality of the translation. Accordingly, AI has a great potential as a support for translators and interpreters: this flexible tool can be adjusted to the needs of a given sector, or even to the needs of particular companies.
One thing is certain: we work on new technologies to serve humanity, and not to hinder its growth. What if translators become revisers of AI-generated texts in the future? What if we no longer have to translate repetitive texts with no literary value – contracts, terms and conditions, manuals…? And what if translation by people, and not by machines, becomes a premium service?
At Alingua, we are prepared to try out new technological solutions and use them to support the daily work of our translators. We don’t fight the inevitable progress. On the contrary: we learn and follow the news while trusting in the timeless value of human creativity. After all, years of knowledge and culture, rich experience, and a personal, emotional view of the world outgrow quickness of learning and processing new information by a lot.