Last week, BBC’s Panorama asked ‘Could A Robot Do My Job?’ According to a study conducted by researchers at Oxford University and Deloitte, the answer is a resounding ‘Yes’ for 35% of current jobs in the UK. It seems that we’re on the verge of a revolution whose impact on jobs and productivity will be equal to that of the nineteenth-century Industrial Revolution. Luddites beware!
Oxford University academics Michael Osborne and Carl Frey calculated how susceptible to automation each job is based on nine key skills required to perform it; social perceptiveness, negotiation, persuasion, assisting and caring for others, originality, fine arts, finger dexterity, manual dexterity, and the need to work in a cramped work space.
There’s a widget on the BBC website which uses these criteria to tell you how likely it is your job it will be mechanised in the next 20 years. Naturally, I was keen to see what my chances were. I selected the author/writer/translator option and was relieved, but not terribly surprised, to be told that there’s only a 33% chance I’ll be replaced by a robot in the future.
The reason, of course, is that most types of writing and translation are highly creative and require a deep understanding of culture, context, and the nuances of human communication in order to be performed effectively. Computers simply do not have these skills and are unlikely to develop them any time soon. They would need to reach full consciousness, as we saw in the hit Channel 4 series Humans, or the 2015 Alex Garland movie, Ex Machina. Thankfully, this is still the stuff of science fiction – and I hope it stays that way. A lot of humans died in the making of those dramas.
Nevertheless, there are still many who believe that translators are about to become redundant, that machines can basically do the job already. These are usually the same people who think that translation is just a matter of substituting the words of one language for another. Most of them have probably never learned to speak a foreign language.
But as any linguist knows, translation is a very complex cognitive operation. There are two basic stages: 1) decoding the meaning of the source text, and 2) re-encoding this meaning in the target language. To accomplish the first, the translator must interpret and analyse all the features of the text, a process that requires in-depth knowledge of the grammar, semantics, syntax, idioms, and other features of the source language, as well as the culture of its speakers. But the translator needs the same in-depth knowledge to re-encode the meaning in the target language, a skill which is often overlooked when the focus is on second-language knowledge.
There are multiple ways in which the same word or phrase can be translated (just think of how many words we can use to say something is attractive to look at). Every choice made by the translator will provide a different tone or register. Some words and expressions don’t even have a direct translation and have to be approximated or rephrased entirely. It is as delicate an operation as writing the original text in the first place – perhaps even more so due to the need to so fully comprehend the purpose of the source text.
Many of us have seen the sometimes hilarious translations thrown up by programmes such as Google Translate. Some people don’t care if the quality of the translation is poor and are unwilling to pay a human to do the job. Others are simply ignorant of the complexity of the task. For the latter type, this can lead to embarrassment and humiliation which they would rather have avoided. Yet more are willing to pay atrociously low rates to unqualified, unskilled translators who are unlikely to deliver quality work, and who may even resort to machine translation themselves.
All this aside, I do believe that as computers become more intelligent certain types of translations will be mechanised. I’m thinking here of the kinds of documents which do not vary much in terms of language or style, for example, birth certificates, user manuals, or university diplomas. And maybe that’s where the 33% comes in.
For the large bulk of translation work, though, us humans have nothing to fear – yet.