Don’t cry for the milkman

Every night when I was growing up, my mother left two empty (glass!) milk bottles at our doorstep, and the next morning by the time we got up, we found two bottles of fresh milk ready for the family’s breakfast. Though it looked like magic to my young eyes, there was nothing supernatural about it. It was the milkman, who every morning delivered the milk on his cart (yes, a cart, pulled by a real horse). Once a month, my father settled the tab, and everybody went about their businesses as usual. Then, one day, the milkman stopped coming and no one took his place. So we started getting our milk from the local supermarket. Milk did not disappear, it was just delivered in a different way.

As the world changes, people have to adapt. Professions come and go – they always have (have you seen a film projectionist lately? was anyone’s grandparent a data scientist or a social media manager?). With the development of artificial intelligence, however, this seems a newly discovered truth. If they haven´t already found out by themselves, people are constantly being told that job security is forever gone, that many of us should seriously consider a career shift. There is hardly a newspaper that hasn´t published an article listing all the jobs that will be lost in the next five or ten years (social Darwinists are rubbing their hands in anticipation). Since the launch of chatGPT in November 2022, apocalyptic trumpets have sounded even louder, announcing the death of some professions. And guess what – human translation is one of them. But is it?

Although machine translation can handle large amounts of texts in seconds at almost no cost, producing for the most part passable or even good translations of straightforward texts (at least to and from widely spoken languages), it struggles with specialised content and text-type specificities, such as legal documents, medical reports and highly technical texts. In addition, MT fails to take into account the cultural context of a text, or the nuances of more creative forms of expression. For these types of translation, human translators will predictably continue to be on demand, particularly highly specialised ones, able to produce prime quality translations.

On the other hand, there is a wide array of other tasks beyond translation proper. Usually frowned upon by old-school translators as ancillary activities, pre- and post-editing and revision, to name but the most usual ones, do require translation competences, as the EMT Competence Framework reminds us. Localisation, project management, intercultural mediation, transcreation, multilingual copywriting are other value-added services that may be provided by a translator.

Shortly put, then, translator training must prepare graduates to become specialised professionals with the skills to, not only collaborate with AI tools, but provide excellent translations. At the same time, those graduates must have the necessary flexibility to offer a complete range of other language-related services. Though it sounds very ambitious, translation programmes in the EMT network do offer students the possibility of acquiring the competences that the market and the profession need. By revising and updating their curricula, listening to the industry, and focusing on analytical and critical thinking abilities, master’s programmes like MTIE succeed in training adequate professionals who encounter almost no obstacles to enter the labour market and build a successful career. And as long as there is a need for accurate communication across linguistic and cultural barriers, professional translators will be necessary.

But maybe there is a threat to the profession coming from a different corner. What if quality ceases to matter? There are signs suggesting that people consume texts (preferably short ones) rather than read them. Internet and social media posts contain appalling examples of bad writing, including grammar and spelling mistakes, and no lexical variety. A growing number of teenagers, especially since the pandemic, have difficulties in understanding complex texts and ideas.  Streaming platforms like Netflix already use the raw output of machine translation for their subtitles, often with nonsensical results, and nobody seems to mind. Even literary translation (the seemingly untouchable sanctuary of creative translation) is no longer safe, with at least one publishing house in Portugal mass producing machine-translated literary classics (see https://comunidadeculturaearte.com/traducoes-por-inteligencia-artificial-ia-chegam-a-portugal-sem-se-fazer-anunciar/). If the day comes when readers cannot tell the difference between a good text and a poor one, why would anyone pay for a good one? And if that happened, translators losing their jobs would be the least of our problems.

Going back to the story of my childhood milkman, we started buying our milk in the local supermarket. Though it was more convenient, perhaps healthier and it lasted longer, at first it tasted funny and we didn’t like it so much. But after a few days, we all got used to it and stopped missing the old one. So please don’t cry for the milkman – for all I know, he sold the farm and retired. It was milk lovers that got worse off.