Machine translation adoption is on the rise - what does this mean for the freelance translator?

Neural machine translation (NMT) is not just a phrase on everyone’s lips anymore- this technology is already being used by “the masses”. Just as computer assisted translation tools and translation memory (TM) technology fundamentally changed translation and led to a significant increase in productivity, NMT is now being accepted as another important new paradigm that leads to higher levels of productivity.

In recent years, the quality of machine translation (MT) has improved quite dramatically with neural MT technology entering the scene – its effects have been felt across the whole of the supply chain. Language service providers (LSPs) and corporate organizations now have the option to translate larger volumes of content than ever before and can afford to translate content that they may have never previously considered. 

For the freelance translator, Post-Editing Machine Translation (PEMT) jobs have become more common, however their perceptions of using machine translation as a productivity tool whilst translating and whether they should accept PEMT jobs are still mixed. Whilst a good quality machine translation provider can aid a translator by increasing their productivity, we often hear that freelance translators are hesitant to utilize this technology, but why?

The challenge for freelance translators

Well for many, finding a good quality MT provider isn’t as easy as it sounds; not all MT providers excel in all languages and MT is not necessarily appropriate for all types of content. There are also challenges around pay, as rates are often significantly cut when MT has been utilised and many translators are concerned that machine translation will ultimately start taking their jobs. 

The results from a recent survey conducted by CSA Research echoed some of these concerns; only 37% felt that the quality of machine translation output they dealt with was good. In addition, 81% of those taking on PEMT work also noted that the raw MT output varied significantly from client to client. If the quality of machine translation is really so different from provider to provider, client to client, then it is understandable why so many freelance translators are reluctant to take on this type of work and why translators are often against utilizing MT as an additional resource whilst translating. 

Why then is MT adoption increasing?

Despite these very valid concerns around using machine translation, we’ve already established that more and more freelance translators are starting to embrace it nonetheless. Of course, when you look at the reasons why, it’s arguably not just attributable to one thing, but many. 

First and foremost, the quality of machine translation has greatly improved over the last decade. With the introduction of neural machine translation, raw MT output is much more fluent than it used to be and handles complex languages much better. Translators are starting to see the benefits of leveraging it and the inspiration it can provide by suggesting terms or phrases that they may not have otherwise thought of. Using MT does not necessarily restrict creativity, but can enhance it.

There are also many different ways you can work with machine translation, either on its own or in combination with a computer-assisted translation (CAT) tool, offering users complete flexibility with how they work. Combining good quality MT with a CAT tool enables users to further increase their productivity and therefore deliver work faster - the faster the job, the more work they can take on. 

In addition, in April 2020 a Translation Technology Insights (TTI) survey was released that yet again solidified our belief that MT is increasing in popularity. In this study, respondents were asked what translation software they planned to invest in next year (that they were not already using) and 50% of the LSPs surveyed stated that machine translation was at the top of their list. LSPs are at the heart of the supply chain, so if they are already embracing machine translation (or at least starting to), it is only natural that their clients and vendors alike are following suit.

So what does the future hold?

As more freelance translators add machine translation to their armoury, the more success stories there will be. The more stories of success there are, the more people will jump on the MT hype - it’s a self-fulfilling cycle. Machine translation is now seen in all levels of the supply chain, so what role does this leave for the human linguist?

With the rate that the world is creating content, it is fair to say that machine translation is here to stay, particularly as it enables companies to translate content that they would never had been able to before, at a cheaper rate. Machine translation is continuously improving; it can now improve more in a few months than older machine translation engines could in years. However, saying that, it is unlikely we will see the massive improvements we have experienced over the last decade- more likely, it will improve incrementally from here. 

Whilst developments in MT technology are sometimes seen as a threat, this can also be seen as good news for the expert linguist. As machine translation improves, the demand will grow for a more ‘expert’ translator. We are more likely to see situations where the basic translation is completed by an MT engine and a specialist translator is called in later to thoroughly review and fine tune the translations. 

As a linguist, this means becoming a specialist in your field is now more important than ever. Those that chose to remain a generalist and accept work from all fields may find it tougher to find work in the long run. On the other hand, the expert linguist will be able to charge more for their time and services and will be able to differentiate themselves from the sea of other translators available.

Translators can also differentiate themselves by adopting new tools and new technologies that will increase their productivity and deliver results quicker to their clients. As we’ve already discussed, embracing machine translation, for example, doesn’t prevent you from being an expert, or even from being creative, it will just enable you to deploy your expertise in a more productive and efficient manner.