Machine translation post-editing: Boon or bane for professional translators and the industry?

Now it comes as no surprise that machine translation post-editing, and machine translation in general, are viewed with suspicion by many translators. After all, history is replete with examples of jobs being lost or made poorer due to technological automation.

Human translation remains a prized gold standard in many applications. And while most translators remain confident about the superiority of human translation, they deplore how machine translation post-editing has become more in demand, ostensibly at the expense of human translation projects.

But contrary to this expectation, we can actually argue that machine translation post-editing benefits translators.

How? By creating a whole new market that needs them.

Human translation is a big investment

Human translation has always been a costly affair, which meant that businesses would invest in translation only after much consideration, and when they are sure they absolutely need it.

There is a story from World War II about engineers trying to design better protection for warplanes. They analyzed planes returning from air raids and took note of where the planes had the most bullet holes. At first, they decided to reinforce these areas to better protect against bullets. But one engineer pointed out that the bullet holes were present on the planes that survived. The ones that didn’t were hit somewhere else—on the areas without the bullet holes.

In the same way, when we look at the demand for translation, we see the “bullet holes” created by more lucrative translation projects because these are the ones that actually make it to the market.

But on the lower end, there is demand for translation that couldn’t be served because of the high price of translation. Often, these are projects that would help with a business’s growth, but not quite enough to commit high resources for translation. And because it couldn’t be served, it remains unseen, and doesn’t “survive”.

The language barrier has always been a major bottleneck for businesses’ growth, and only those with the capacity to invest in translation are able to achieve that growth. And this growth is far from linear! Global expansion is often an accelerating process, as opening up to a completely new market often means greater returns that can be reinvested into further expansion—and further translation.

MT post-editing: a viable solution?

So where does machine translation post-editing, also known as MTPE, fit into the picture?

This is a question that would probably have a different answer if you asked it over five years ago. Back then, the quality of machine translation output was simply not viable enough for use in post-editing. Often the text would be hard to decipher, and it would be just as fast, if not faster, to translate the text from scratch. As such, the job wouldn’t be worth it for any translator, especially at lower pay.

But with the advent of neural machine translation in the mid-late 2010s, MT became much better. The raw output of machine translation has become more accurate and more understandable, often requiring only light edits, if at all. Now, translators are able to look over a text and work on it much more quickly than before.

Learn more here: Everything You Need To Know About Neural Machine Translation

With machine translation post-editing becoming a more feasible solution, translators have become able to provide language services at lower cost than before. But isn’t that a bad thing? Not necessarily. Let’s go back to the story about the planes. Before MTPE, there was a lot of demand on the lower end of the spectrum that couldn’t be fulfilled because the price couldn’t justify them just yet. But with MTPE, translation is suddenly accessible, and this previously unseen demand is served.

MTPE creates opportunities for growth

Which means there are more planes surviving—and more jobs for translators across the board.

The demand for human translators doesn’t change much with the rise of machine translation, especially on the upper end of the spectrum. For one, there are many cases where it is crucial that language is absolutely accurate, such as in law or healthcare. There are also cases where creativity in language remains an important aspect, such as in marketing or literature. These are cases where human translation remains in high demand.

But now, there are many new use cases, particularly in business, where machine translation post-editing is enough. This is often the case with texts that are high in volume, require quick turnaround, and are of middle value. These are use cases where the cost of human translation would normally be a barrier to entry, but with MTPE they become possible.

This advantage to businesses isn’t one-sided either. Translators, and the translation industry as a whole, stand to benefit.

The translation industry relies on every other industry’s need for growth in order to maintain itself. What drives demand for human translation is a healthy business landscape. In particular, it relies on businesses that can afford translation.

MTPE gives more businesses access to global growth. It gives them that initial boost that helps them break through the language barrier toward accelerating growth, which means more businesses are there to generate greater and greater demand for translation.

Responsibility and ethics in machine translation post-editing

Which isn’t to say that everything’s good. There’s much that businesses—and even language service companies—need to understand in order to foster a good relationship with translation professionals where MT post-editing is concerned.

For one, it’s important to remember that not all machine translation is equal. There are some languages where the quality of output is good, and many others where it isn’t so good yet. In the latter case, MTPE may not be a good solution, as it puts the translator at a disadvantage.

Next, it should also be understood that while MTPE increases the translator’s efficiency at working with a text, there’s still a limit to what can be humanly done, and clients should set expectations accordingly. That way, translators would be less wary of taking on post-editing projects.

Finally, clients need to know what to expect when it comes to machine translation post-editing. For one, it’s unreasonable to hold it to the same standard as human translation. MTPE also comes with other risks that clients need to be made aware of, so they can make an informed decision about whether they need MTPE or full human translation.

In all of these cases, it’s often the role of the language service provider to mediate between the client and the translation professional. As such, they hold the responsibility of advising the client and helping them choose the right solution, while also upholding the welfare and dignity of the translator.

To sum it up

Machine translation post-editing, and machine translation in general, have opened up a world of possibilities that can help both businesses and translation professionals flourish. In the bigger picture, both sides stand to benefit from the accessibility that machine translation post-editing provides. But there is much that needs to be taken into consideration in ensuring that the growth and development they bring are sustainable.