There is no doubt about it. Machine translation is getting better and better. There was a time, not so long ago, when you could easily tell the difference between a machine translated piece of text and one that had been professionally translated by a human translator. That difference is getting narrower as the technology improves. It would be fair to say that machine translation, or MT for short,is already matching the level of an amateur translator, even if it is yet to get to the level of a professional.
Success feeds on itself and the fact that machine translation has improved means that more people are turning to it, even if it isn’t being used widely where quality really counts. Increase in demand feeds a research and development effort which continues to search for refinements in the technology, meaning that the world can expect even greater improvements in the future. But will freely available machine translation technology ever be a substitute for the human translator?
The short answer is probably negative, at least in the shorter term. The amount of development into MT depends partly on how much it is worth developers to spend time on fine tuning the translation technology. MT, in other words, like any technology, has to be worth developing in simple economic terms. The experience is that there are now sufficiently large numbers of people who are prepared to use MT to make the search for better technology worth it. While MT is unlikely to replace the human translator entirely, it can certainly cut down the time taken for translation, even for the professional. In some cases, using MT to do most of the work with an important emphasis on post-editing by the human translator may soon become for the norm for bulk, repetitive and non-critical material.
One negative aspect of the improvements in MT is its widespread uptake by translation agencies that are not so professional. The internet can become a dumping ground for both the good and the bad, the useful and the not so useful, so the discerning client who is looking for value for money in terms of translation quality has to be much more careful about who they pick for their translation work. The lesson is that as machine translation forges ahead in leaps and bounds, it will ultimately improve translation quality as long as it is used wisely, but customers – and translators themselves – do need to take more care about how it is used and who is using it.