The first thing I hear whenever I tell someone I’m a translator is: “Don’t you worry that your profession will soon disappear because of technological advances in machine translation and Artificial Intelligence?” This is typical of laymen who think that the only thing translators do is “translating”.
Today, in the last article of this series celebrating my 25 years in the translation industry, I will answer that question and write down a few considerations on how I see the future of human translators.
Table of Contents
Human translators and technology advances since 1993
First of all, let’s do a brief comparison: What has changed in my profession between 1993 (the year I started my career as a human translator) and today? Without a doubt, my answer is: today, the translating profession is more heavily driven by technology tools and the digital transformation.
In 1993, to research specialised terms for my translations, I used to visit several university libraries and had to laboriously write down everything (at the time, notebooks were very expensive) before using the data in the actual translations. Today, terminology research takes place in the blink of an eye with the help of Google and my large database of documents I translated in the past 25 years of professional experience.
Professional translation software
In addition, all my translations are processed with a professional translation management programme, not simply in Microsoft Word, as I did when I started translating professionally (let alone with a typing machine or with handwriting, as I did during my translation education in the late 1980s). The development in professional translation software is fast-paced; new features are frequently added with the goal of translating even faster, more productive and more quality-oriented.
You can’t dissociate professional translations from such tools anymore. The core skills of a human translator are still there, of course, but the machine has become an inseparable companion. I think what’s happening in the translation world is not dissimilar to the situation in, for instance, architecture or engineering.
Technological skills
The technological advances in the translating profession also means that freelance translators must acquire more and more skills and knowledge tools that weren’t required 25 years ago, such as:
- knowledge of HTML, CSS, PHP, XML and Java for securely handling tags in localisation projects, but also for creating or at least maintaining a website according to current standards;
- online marketing, including content marketing strategies to attract new customers or at least perform public relations through social media;
- using integrated office management systems to quickly generate price quotations and invoices, and to manage customer data and tax information;
- knowledge of information security measures to protect against cyber attacks;
- as well as systematically following up on industry developments through digital publications.
Translation, a highly competitive market
The advances in technology tools for human translators have reflected in market changes as well: the translation market has always been highly competitive for the most common language pairs, but the digital transformation has intensified competition even more considerably by enabling an increase in workforce availability. The greater the number of available human translators for a certain language pair, the more difficult it becomes for new translators to set a foothold in the translation market.
As a result, many translators are competing with one another purely based on price. And where there are clients who consider the price to be the main decision factor when hiring a translator, a vicious circle is created that makes it impossible for many young translators to set themselves free.
After all, it’s naïve to offer low fees in the beginning and hope to “increase them gradually”, as customers are not stupid and will look for a cheaper, probably novice translator as soon as their current translation provider asks for higher fees.
Translators who prefer to charge cheap rates instead of investing in specialisation and adding value to their services don’t even realise that when doing so they are harming not only their competitors but also, in the long run, their own business and the translation industry as whole.
Human Parity in machine translation
The success of machine translation (MT) – something that has been repeatedly praised by the media in recent years – is based on the concept of human parity, i.e. on the perceived quality of machine-translated texts when compared to human translations. With every new machine translation technology – from rules-based to statistical to neural MT -, translation quality is perceived to be higher and higher, which, in turn, generates expectations of bigger and bigger leaps.
However, quality leaps haven’t been big enough to achieve perfect parity. For instance, the MT system developed at DeepL is known to generate translations with almost no grammar mistakes, but which don’t always convey the correct meaning.
Translations with Artificial Intelligence
Even the latest Artificial Intelligence applications have not yet reached the quality level of a human translation. For instance, it is known that AI-generated texts contain lies or at least factual inconsistencies that can only be detected at further analysis.
Laymen may think that machine translation are good – but only because they’re laymen, don’t have deep knowledge of the foreign language in question and haven’t been trained to perform further analysis of such texts and the underlying technology (i.e. they lack digital literacy). This, of course, is very dangerous and is happening right now not only in the translation industry, but also in many other areas where AI is being used.
In addition, the quality of machine translations is highly dependent on the language pair: German-to-English machine translations are considerably better than German-to-Portuguese translations. Language pairs for which large amounts of data (and translations!) are available can be better processed by MT systems. This means that the quality of future machine translations will also depend very much on the quality of existing human translations.
Therefore, in translation, humans still set the standards with which to measure the machines’ success!
Specialised human translators
Let’s suppose that the quality of machine translations will become higher and higher for more and more language pairs, not just for the most common languages. In this case, human translators will indeed become redundant.
On the other hand, a human translator may not be needed for all kinds of translations.
For example: A text from the internet serving as a quick source of information doesn’t necessarily need to be translated by a professional human translator – Google Translate or DeepL will be good enough.
But:
An important contract containing sensitive information will be better taken care of by a specialised legal translator. The same applies to court documents as part of a legal case. And to all kinds of texts that are strategically important, contain personal data or need to be published.
Therefore, despite all technological advances, specialised human translators will still be needed.
What do human translators do that machines don’t?
Those who ask if human translators still have a future don’t understand – because they are laymen – a fundamental fact:
Professional human translators – that is, qualified translators who earn a living from translation instead of translating as a side job – do so much more than simply “replacing a word in a language with a word in another language“.
From the perspective of some who has been working in the translation industry since 1993, I can safely say: in all these years my work has never been about “just translating”.
Here are some examples straight from my professional practice:
- As a professional financial translator I find factual errors – and of course typos – in financial reports and tell about them to my clients, so the texts can be edited before publication, thereby avoiding additional costs for corrections and republication.
- As a sworn legal translator I not only assume liability for the accuracy of my translations but also advise my clients about which documents are needed and which ones aren’t in certain bureaucratic procedures – always based on many years of experience with German-Brazilian bureaucracy –, so that they save time, stress and above all money.
- As a website localisation specialist I tell my clients, when necessary, that a certain sentence or product would be inappropriate in the target country (in this case: Brazil) because of intercultural differences, so that the offering can be adapted before launching a marketing campaign in the that country.
All of that can only be offered by a human translator – not by the translation system of some AI app.
Why hire a human translator?
For the masses who need translated texts merely for informational purposes with no special requirements at all, professional human translators are simply too expensive, slow and complicated. And there is nothing wrong about that.
After all, you don’t hire a doctor or a notary public “just like that”, but because you need their professional services. Same thing with professional human translators.
Professional human translators are really only needed when an especially high degree of quality, expertise, confidentiality and seriousness is wanted – that is, whenever a human being with qualifications is needed.
Will the translating profession soon disappear?
To answer the question posed in the beginning of this article, as a specialised human translator, I don’t worry about the future of the translating profession in the face of technological advances in machine translation and Artificial Intelligence.
What I find more problematic and threatening is the fact that young, qualified colleagues are selling themselves below their worth and thereby contributing to the downward pressure on prices in the translation industry as a whole. After all, a known fact is that, if you’re cheap today, tomorrow you’ll be competing with someone who is even cheaper and who will in turn one day become too expensive, and so on and so forth. This is the good old race to the bottom.
To sum up, I’m not afraid of translation machines; after all, as a professional translator I’ve learned to master productivity-enhancing working tools.
I also don’t fear merely bilingual people who work (or want to work) as translators without translation qualifications; the market will eventually separate the wheat from the chaff.
What I do fear is qualified colleagues who don’t understand that they can offer a unique selling point as qualified human translators provided they invest in continuing professional development and specialisation.
If the translating profession disappears one day, it’s translators themselves who are to blame – not the machines. Until that happens, I prefer to focus on assisting demanding customers on the basis of my specialised skills and many years of experience as a human translator – distancing myself from the bulk market of anonymous vendors. This is the exact same strategy, by the way, that has always kept me busy with translation work throughout the years.
[With this article, a long series celebrating my 25th anniversary as a professional translator comes to an end. I had fun writing down memories, insights and observations and sharing a few discussions about the past and future of the translating profession. If you missed the other articles, here’s a useful overview.]
