With AI having taken the world by storm since the November 2022 emergence of ChatGPT, many translators have seen their work dry up—or have been offered pennies on the dollar. Some have left the profession, while others are seriously (secretly?) considering it.
So, this seems as good a time as any to ponder the future of our beloved profession.
Let’s start with a few basics.
After a slow start in the mid-2010s, neural machine translation (NMT) has become widely used in the drafting stage. In fact, 98% of surveyed translation agencies use NMT (see the recent DeepL report on this topic), and freelance translators often do so behind the scenes to remain competitive.
In terms of quality, a few years ago, researchers found that, during post-editing, less than one in five segments required no changes (see, for example, Evaluating the usefulness of neural machine translation for the Polish translators in the European Commission by Karolina Stefaniak, 2020). Today, NMT engines have improved by leaps and bounds (I know because I’ve compared machine outputs from a few years ago with what the engine gives me today). Anecdotally, in 2024, perhaps 25–30% of segments don’t require any human intervention, at least in certain “predictable” text types such as policies. Of course, this is only a general number, and it will depend on the language pair, tool, and specialization.
But still. If graphic designers could basically get one-third of their work done for them upfront, even just some of the time, that would be a game-changer, wouldn’t it?
This is a big deal, and we shouldn’t downplay it.
Compounding this state of affairs is genAI, which is able to produce roughly the same quality output as NMT by leveraging a pretrained model to produce human-like responses. It has different strengths and weaknesses than NMT, which actually makes it highly useful and complementary.
This can be seen in certain boundary-pushing tools such as Bureau Works, which has genAI clean up the draft NMT translation and factor in the translator’s own preferences right in the first draft. The time savings can be stupendous.
The jury is in: automatic translation tools are improving at breakneck speeds.
For most professions, this isn’t a problem—it’s a win-win. Just think of an architect using some AI-powered AutoCAD. The client gets their desired blueprint faster and the service provider is able to make more money in less time (possibly with other benefits, such as automatically taking into account municipal by-laws to avoid permit problems down the road).
In translation, the portrait is more nuanced. Certain business-savvy translators are able to turn AI into a plus for their clients, adjust their rates like everyone else in today’s economy, and make a great living. But others are seeing devaluation and price stagnation even as their work volume plummets (with some clients finding “in-house solutions” to their translation needs).
Furthermore, anyone can play around with DeepL and ChatGPT, so our work instruments are often more democratized than, say, a highly specialized AutoCAD program. This doesn’t help clients see the value we bring.
While I have no crystal ball, I do think that, taken together, these considerations suggest an overall futurescape that I’m going to attempt to describe below.
Translation programs are closing as prospective students (and their parents!) hesitate to register for university translation programs in an AI-dominated world.
Not all programs… but enough that this doesn’t feel like a temporary or regional issue.
The situation is fairly widespread, at least in North America and Europe.
Concurrently, translators are leaving the profession in droves out of fear for their financial security. And again, who can blame them?
Finally, it’s important to know that the average age of translators is generally much higher than in other professions. Last I checked, in Quebec, the average age of translators was close to 50—while the average for all professions combined is usually around 35!
That means lots of translators are going to be retiring in the coming years.
When you put these three facts together, it doesn’t take future vision to see that there will likely be fewer translators moving forward.
And this may be a huge boon, because the translators who stay might fare really well in the coming decade or two. Perhaps, with AI, there will be fewer translators who process larger volumes of text, and the higher demand will support higher rates (for the few who dare to raise them in order to stay in business).
The mid-market has been shrinking for several years.
One result of this is that translators today often fall into one of two camps. Either they are struggling more and more, or they have gone premium and are increasingly prosperous.
What makes the difference? Some of it has to do with specialization.
Generalist translators and translators working in feast-or-famine fields are sometimes having a hard time getting the rates and volume they need to thrive.
Meanwhile, highly qualified translators working in specialized fields have less and less competition with no attendant decrease in demand. They’re doing well!
The price surveys I’ve come across in recent years tell of stagnating rates. But I suspect that this is only a surface-level view of what is really happening. Behind the scenes, some translators are earning a lot less, and some are making a lot more.
This is what translators have been talking about on social media and professional forums. Many excellent translators just happen to be generalists or to work in more precarious fields, and today, they are really struggling, which is unfortunate. But there is a solution, and that brings me to my third point.
“I never had to market my services before. I got plenty of work through word-of-mouth and referrals. But something has changed since 2022. The work has dried up and I think I need to learn marketing (even though I absolutely hate it)! Help!”
Does this sound like you?
If so, I want you to know that you’re not alone.
Most of us became translators out of a love of words and language (not to mention a desire to help our clients and bridge cultural divides). We’re introverts in an extroverted world. We’re not marketing people, nor do we want to be. Am I right?
But here’s the thing.
One way that translators today are thriving is by redefining their role and better showcasing their value to clients, and that requires some marketing savvy.
All the expertise in the world will get you nowhere if your clients are not aware of it.
A good example is translators who are picking up new skills (for example, in SEO translation) and recasting themselves as “language solution architects” or “your language partner.” They’re able to offer turnkey service, make a compelling case for their added value, and charge accordingly.
And they are killing it.
The fact is that many (most?) of us translators have survived off referrals for decades, but in a changing world, we will have to evolve in order to get clients to understand all the value we are bringing to the table.
In my humble opinion, this “businessization” of translation was long overdue. If you’re a freelance translator, you are also a business owner!
As a profession, in order to get clients to see our value, we are going to have to learn business.
The good news is that other professions went through this process decades ago, and we can learn from their practices and adapt them to our own field. In fact, that is exactly what I have been working on and teaching through LION Translation Academy and my own business mentoring.
Well, it depends.
If we collectively bury our heads in the sand, continue to charge the same rates as we did 10 or 20 years ago, fail to specialize, work in silos, and hope for referrals, we may indeed find our profession grinding to a halt.
Markets can collapse.
Conversely, if we become qualified in one or two in-demand specializations, partner up with other professionals, accept the possibility of learning complementary skills, raise our rates to continue to serve our clients well, and (most importantly) master the art of marketing, those of us who remain in this field have a bright future ahead of us.
I have zero doubt about it, because many of us are already thriving in this new era.
To come back to what we were saying about machine translation, remember that at the end of the day, NMT does nothing but recognize patterns, and genAI merely spits out human-like responses.
The same human qualities that machines appear to devalue (requiring time to think, for example, or putting emotion and authenticity into our clients’ texts) will save us.
But only if we develop the expertise to make our clients aware of the value we are providing in order to help them succeed.