As machine learning and artificial intelligence becomes more and more prevalent, many careers are now becoming obsolete or decreasing in size. For the translation industry the rise of more advanced technology seems to pose a threat to many jobs; however, will it ever be possible for the industry to be completely dominated by machine learning? Here are some reasons why machine learning will ultimately never replace translators but will be a tool that many translators use to their advantage.
Many companies such as Iflexion, Hidden Brains and Icreon are serious about implementing machine learning. Machine learning has made many leaps within a short amount of time; nevertheless, all machine learning is confined to specific rules created by programmers. This means large amounts of data can be collected and used to make predictions by the machine; however, it cannot organically learn new phrases or make creative decisions. Therefore, rules within a language or small grammatical difference between languages which can be irregular, will often be overlooked because they appear anomalous within the data collected by the machine.
In essence, it can only produce what humans have already done and follow a pattern. However, this will cause the loss of nuance in many phrases within the language. Many of the finer details are often lost in translations by machines because language is influenced by culture and society. Machine learning relies on rules and guidelines to function appropriately; in contrast, translators bend the rules, adapt the semantics of a phrase and use completely new analogies to achieve the same point. This process cannot be achieved through rules or automation.
Although, artificial intelligence and machine learning will continue to improve and upgrade, a machine will not be able to combine creativity, culture and morality into their translation process. So how will machine learning help translators? Despite not being able to translate all details and nuances, machines do a good job at translating parts of a language where a lot of data has been collected. Effectively, translators will be able to use machines to help them translate parts of a text and then do what algorithms cannot do… be creative.