The process of digital adoption in the translation industry
After a considerable amount of time in the shadows, the translation industry has moved into the global spotlight over recent years. The world has witnessed and taken note of the exponential growth of generative AI tools (set to become a $1.3 trillion market by 2032) and their ability to refine translation tools and software.
With 72% of Fortune 500 Chief Human Resources Officers believing that AI will replace jobs in their organisation in the next three years, it looks like we’re set for change.
The question is, can we in the translation industry adapt and learn to love these new developments?
Index
In this article, I’ll cover…
- What is digital adoption in translation?
- A brief history of digital adoption in translation
- (Fairly) recent tech developments in the translation industry
- How quickly have we adopted these translation technologies?
- The future of digital adoption in the translation industry
What is digital adoption in translation?
There comes a point (for most of us anyway), where we can start to feel a little bit lost. Right now, at least to me, it feels like I’m being swept up in this wave of doomsday talk:
“AI is taking over”
“Machines are flooding our profession”
“We’ll end up as mindless, underpaid machine translation prompters”
“We’re losing our value, and we’re losing it fast”
What choice do we have but to sink or swim?
In short, digital adoption is when all available tools and their features are being used and leveraged to their full extent. As someone who’s been working for Language Service Providers for a long time, I’ve always stayed on the pulse of technological development in our industry. There’s been a lot. Our everyday working lives are completely different now compared to even 10 years ago. But I think the good news is that (whether it was our choice or not), we’ve coped with and adapted to these major changes pretty well.
Let’s go back through the most significant digital adoptions in the translation industry and ask ourselves: is the onset of AI-assisted technological developments a major shift in a different direction, or is it the same old situation, playing out over again?
A brief history of digital adoption in translation
Only a decade ago, I was working as an Assistant Project Manager in a small translation agency. It was a shabby little office. The stained blue office carpet had seen better days. I say blue, but it was actually a faded pinkish-purple colour in the spots where the light hit through the dingy windows. A constant draught flowed through the sash windows, and I stashed a thick woolly cardigan in my desk drawer to wear over my clothes throughout October until May. The office was always eerily quiet, with the exception of several furiously tapping keyboards, the occasional creak of a chair when someone shuffled their bottom (presumably to prevent it from going numb from all the sitting down), and of course the sporadic bleating of the phone.
One day, sat frozen in the middle of trying to organise a complicated translation project, I heard a sudden crashing sound accompanied by a dial tone that sounded like nails on a chalkboard in the corner. I sprung up to investigate, and found a fax machine buried under a pile of folders and dust. And that’s when one of my colleagues piped up “Cor, we haven’t heard that in a while!”.
Clueless, I asked them what it was there for. To send and deliver files of course. A relic, the former hub of the office, sat there unused for several years, before the widespread adoption of email. Until that day when a random junk message rattled through its paper tray.
The fax machine wasn’t the only thing we left behind. We used to have to hand count scanned handwriting or dodgy PDFs, as we didn’t have reliable OCR software available to help us with it. We used to split the batches of paper between us, and I was always the slowest counter. When I said this out loud, my colleagues told me “Well of course, you’ve had less practice. We used to have to do it this way all the time”.
Nowadays, it’s pretty normal for a client to expect a 10k page report to get translated within days. That wasn’t always the case. Years ago, it would have been translated using a typewriter and sent via post or courier. Things are much faster now, and thank goodness for that.
And of course, there would have been something else before then. Before fax machines existed, before typewriters were used, before Gutenberg put out his printing press in 1454 even. Translation is one of the oldest professions in the world. If we hadn’t started adopting technology, we’d still be carving out symbols on clay tablets.
But I’m going to fast-forward a bit now and start talking about the more recent changes I’ve seen during my time working in the translation industry.
(Fairly) recent tech developments in the translation industry
CAT tools
Let’s start with Computer-aided translation (CAT) tools. Software programmes designed to speed up the human translation process. By splitting documents into segments and applying a Translation Memory and a Glossary, you can leverage existing translations from Translation Memories, and reuse them.
CAT tools have been around for years, but only more recently have they become an essential part of a translator’s toolkit. In 2023, a survey conducted by ProZ showed that 93% of full-time professional translators used at least one CAT tool. But here’s the thing: the same survey found that only 10.4% have been using one for over 20 years. That’s some serious translation tool uptake.
Translation Management Systems
I was around for the rise of the Translation Management System (TMS), and I can tell you that when I tried to get our regular team of translators to work in one, I was met with absolute mutiny.
In a nutshell, a TMS is a cloud-based platform that combines CAT tools, machine translation, and project management features. For a project manager like me, they were a godsend. Transforming our efficiency, and helping us drastically speed up our process – something that our clients desperately wanted.
It’s been about eight years now since many agencies like the one I worked at started adopting them, and they seem to be much more widely accepted now. At the time, some of my best translators didn’t want any part in it, having shelled out for their own CAT tool themselves, or finding them too difficult to pick up. I used to have to find a workaround so that they could still do the job, but I haven’t encountered this scenario for a good few years now. Like it or not, it’s kind of widely accepted that if you want to work for an agency, you’re going to have to use their chosen tool.
Machine Translation
Machine Translation used to be the butt of our jokes. The thought of using it in a professional situation used to be laughable. We used to secure clients on the promise that we didn’t, and would never use it.
Because it produced poor results. But then after a while it got to the stage where this was no longer a good thing. Our clients wanted the option to have quick, rough and ready translations when budgets were limited and time was tight. When we couldn’t provide that, they started to drift away from us.
But that’s probably because this all coincided with Google Translate’s announcement that they were switching to a neural machine translation system. By contrast to rule-based or statistical machine translation, neural machine translation took the whole input sentence into account when generating the output (rather than mapping individual words or phrases). In other words, it got cleverer.
And it’s still getting cleverer.
Machine Translation Post Editing (MTPE)
And as they got more and more clever, collaborations between humans and machines became more and more common. Humans (or should I say professional linguists) can now edit AI-supported MT to produce a good quality translation. With their expert cultural knowledge and linguistic skills, translators are able to correct any mistakes and inconsistencies that the MT throws up.
It takes skill, and in order to make it financially worth your while, you have to become accustomed to a certain way of working: the end result should be accurate and functional, but there’s no time for carefully crafted words.
So far, from what I can see, not all human translators are loving MTPE work, with some opting to leave the profession altogether. The underlying feeling is that the work is not as challenging, interesting, or financially viable.
How quickly have we adopted these translation technologies?
As Renato Beninatto points out, the adoption of technology is influenced by industry-specific needs, with legal and regulatory framework having an effect on technology adoption, even when new advancements are available.
Technology doesn’t become embedded into a system overnight, it’s more evolutionary than that. You might be surprised to learn that the first machine translation tools appeared in the 1950s, and the first commercial CAT tools began to appear during the 1990s.
It took a while for these technologies to become fully adopted and embraced by the translation industry, and even in the 2010s it wasn’t uncommon for me to encounter a translator who didn’t use one. Perhaps that’s why more recent developments seem a bit scary. Alarmingly, they’re coming into the fore at a much quicker rate.
The future of digital adoption in the translation industry
The global translation services market is still growing, with Nimdzi estimating that the industry should reach $90.8 billion by 2027 (from an estimated $69.3 billion in 2023). But the industry is becoming more competitive than ever, and LSPs are leaning in on AI-enhanced machine translation and processes to improve their productivity and speed, and help them undercut their competitors. The knock-on effect is felt lower down in the chain.
Could this somewhat hostile reaction from translators towards emerging translation processes like MTPE be like all of the other times, a stumbling block on the way to fully adopting this new technology? I’m not so sure.
While I’ve seen the same kind of reactions before when it comes to adopting new technology like CAT tools and TMS, I’ve never seen it to this extent. I’ve no doubt that there are exciting times ahead in some ways, with many new education and reskilling opportunities that await us. But this time, I’m questioning whether it’s us translators, the beating heart of the translation industry, that will be the ones who truly benefit from this change.
What do you think? Are you embracing the adoption of new digital technologies in the translation industry, or feeling sceptical?
Additional references
https://www.gartner.com/en/doc/emerging-technologies-and-trends-impact-radar-excerpt