Post-editing is not translation.

[I have retired from efforts to help translators survive (it’s basically a fruitless activity), and I’ve noted the reasons elsewhere. Idiocy is being peddled as hope, and I won’t be a part of it.]

I notice that numerous translation sellers offering “MTPE work” on their click-work platforms or on commercially provided reverse-auction platforms such as ProZ are attempting to categorize post-editing as “translation” or saying they’re looking for translators. This was sometimes the case on the JAT Job Board, but that might have been because editing or post-editing cannot be selected when you post a job and, to be fair, some said “Translation (MTPE),” but that is still incorrect, considering the task being described and the payment being assumed acceptable.

This looks like just the next step in agencies grooming their essentially captive labor supply to accept the new order.

Although I’ve never done or been asked to do post-editing of machine-translated output, I hear many people say it requires the skill of a translator. I can certainly believe that, and that might be even more the case recently, given the deceptively human-like English quality produced by AI.

As a task, however, post-editing is not translation, and it doesn’t come with an earning potential anywhere near what performing the task of translation provided to freelancers in the past.

Post-editing is no more translation than editing and rewriting Japanese-to-English translations decades ago was translation. Many people doing it then, however, aspired to become translators who actually get to do translation and be paid for translation, and many did achieve that goal. That will not be the case with the vast majority of people who do post-editing now, either to survive mid-career or because they think it’s a stepping stone to translation. It is not, and for most translators post-editing is at best a survival job for people not in a position to reinvent themselves.

The places needing post-editing realize they don’t need to provide much earning potential because of the continued availability of a captive labor supply with not many options. If they did need to provide anywhere near the earning potential of the past, there would be not much reason for them to adopt AI.

Because of the age demographics of the translator population, this situation should continue for at least a decade or two; after that, it’s difficult to tell where things will go, but it won’t matter to mid-career translators currently facing this situation, because for them the possibility of translating for a living will have ended.

All of that uncomfortable reality aside, I wish people would stop calling post-editing of MT/AI output translation, and also wish they would realize that calling a post-editor or even a translator a “linguist” has never put money in their bank account, and that most translators would rather have the money and the job satisfaction than the bloated title.

A Non-Crisis

There has been a kerfuffle recently in Japan surrounding a term that can be translated as existential crisis. It revolved around the newly elected PM’s mentioning that an attack by China on Taiwan could be taken to be an existential crisis, apparently for Japan, although that was not explicitly stated. This led me to think about something closer to the lives of translators that could be called an existential non-crisis.

Some translators might think that the appearance of AI has created an existential crisis for human translators. I disagree, because the use of the term crisis implies a possibility of surviving the crisis. For almost all freelance translators—and arguably many translation companies, although a bit later—survival as a translation provider isn’t in the cards.

It is time for both translators and their organizations to stop pretending—either by studied silence or by active denial and diversionary handwaving—that human translators are not already being replaced by AI on a devastating scale or won’t shortly be almost totally eliminated from the translation process step.

It is also time to realize that the essential and immutable enabling condition for the demise of human translators by AI use is the two-tier brokered structure of freelance translating, and that this condition was satisfied decades before AI appeared. That structure welcomed and fostered a population of freelancers who are willing to be isolated from translation consumers—some even welcomed that isolation—and who are incapable of surviving when the only client demographic available to most of them decides to replace them with AI.

Continuing the pretense of resilience in the face of the AI transformation might feel good (or at least might feel less painful than the truth), but it won’t change things.

Let’s get real and look at the real future, which will be radically different from what most translators imagined when they started translating. For most current freelance translators, that future won’t involve translating.

Are translators’ organizations giving this any thought? It’s hard to tell. One organization I am familiar with is basically silent on this situation. They hold occasional events on topics of interest to a small group of translators in a limited subject-matter domain, but they appear to have lost relevance to many members, judging from their significant loss of members, starting in 2016, long before the pandemic, which has sometimes been cited as the cause of the drop in membership.

I placed this post here because I suspected that, placed somewhere else, it would evoke serious pushback, denial, and even anger from the very people who will shortly be—or already are—ex-translators. Pushback, denial, and anger won’t change things.