AI in Translation and Localisation in 2024

An Independent Assistant or a Human-Supervised Tool?

Reflecting on 2023, many were the trends that left their mark on the translation and localisation industry. From multilingual SEO, Large Language Models, to an increased demand for localisation and the further application of AI and machine translation in the industry, 2023 was a key year with regards to many developments and challenges. From larger enterprises to independent, freelance linguists, many are the professionals who see AI and machine translation as not simply a hype, but as tools that are here to stay in 2024. In what ways will machine translation prove useful, and what are its current limitations? Could AI, as some claim, lead the translator to go extinct? To find out, read on.

AI and Machine Translation in Application: A Helping Hand – or is it?

Research in localisation has demonstrated that AI can be successfully applied on an already known document. It can check the document for spelling, syntax, grammar and single-character errors, provide state-of-the-art speech recognition and translate in up to 100 languages. This allows the professional to apply a quick, efficient, and accurate translation without losing any of the original text’s tone and expressions. Nevertheless, and despite AI’s undeniable positive impact on our work life, it is also necessary to be aware of the dangers and challenges that come with it.

Only recently, in September 2023, the use of AI made headlines in New York. More specifically, the city’s Mayor, Eric Adams, was called out for using audio deepfake in a number of languages he could not even speak! Among these languages, all of which are spoken in Adams’ constituency, were Spanish, Cantonese, Haitian Creole, Mandarin, and Yiddish. As Adams argued, this was his sincere attempt as a politician to speak to and connect with voters in their language. In many ways, Adams’ use of AI was his way to appear more humane and relatable. Perhaps the human touch is what we need, after all.

The Expert-in-the-Loop as AI’s Manager: A Hybrid Solution

Human aspect

Various reputable specialists and researchers, similarly, highlight the dangers of leaving machine translation and AI tools to their own devices, which usually happens when companies are looking for ways to save money. They point out that, due to AI limitations, many low-resource languages face the danger of the final translated text being burdened with a great number of errors.

As you may have already guessed, the solution, many specialists are now claiming, lies in the expert-in-the-loop; in other words, the expert linguist, or post-editor who manages the machine translation output. Indeed, the Chartered Institute of Linguists (CIOL) CEO, John Worne predicts that ‘as we move forward, the ability to ‘ask’ and ‘task’ AI will become an essential skill for professionals, and aptitude in and with languages will play a crucial role in this. Generative AI is now demonstrating to us all that languages are the ultimate human ‘meta skill’, not just a means of communication’. In essence, human oversight and machine management are predicted to be indispensable to the successful end result in translation and localisation processes for 2024 – and we couldn’t agree more.

Why the Human Translator is Irreplaceable

human and ai hybrid

When in need to remind ourselves that AI has still got a long way to go before it can catch up with the human mind’s sophisticated ability to decipher emotions through translation, we need only take a look at literature, also known as the last bastion of human translation. In fact, neural machine-translation can only successfully translate about 30 percent of novel excerpts – all of which are usually simple passages which only require a literal translation.

Machine translation begins to particularly struggle when there is no perfect way to translate the phrase or text. Human translators, on the contrary, can boast being able to choose between a multitude of options in the target language. As they know, the solution doesn’t always lie in the literal translation, which an AI tool may opt for, but rather in the choice that will work for the greater good of the piece – and if you think that this applies only to the world of belles-lettres, think again.

Take a look, for instance, at the film and dubbing industry. Here, the expert needs to manipulate the text and ‘prompt’ the voice engine to achieve the most natural sounding synthetic voice, including prosody and emotion. The synthesized voices are simply not able to pick up nuanced and complex human emotions without being managed by a human expert. In essence, the tone, the soul, and the sensibility of creative copy can only be successfully adapted to a target language through critical approximation, prioritisation and creativity.

It is this creativity that machine translation lacks and it is with this thought in mind that we venture into 2024. AI is and will continue to evolve into an extremely useful helping hand in the industry in 2024 but, as researchers and professional bodies confirm, there is still a long way to go for machine translation to be considered fully trustworthy and efficient.

As Anthony Burgess, author of A Clockwork Orange, once stated, ‘translation is not a matter of words only: it is a matter of making intelligible a whole culture’. In other words, when dealing with the nuances of the human mind, another human mind, or an expert-in-the-loop, is irreplaceable, and will remain so for years to come.

If you are interested in using AI for your content and translation projects, talk to us.

Get In Touch

We speak your language

Get in touch
Get In Touch