machine translation

Mar 05 2024
Corinne McKay

The real question to ask about AI/MT

Greetings, Training for Translators readers! I hope you’re all doing well, and happy March!

First off, I wanted to alert you to a Chinese<>English interpreting class being offered by my friend and colleague Rony Gao. You cannot go wrong with Rony’s offerings, and he’s leading a Chinese<>English consecutive interpreting class, online, starting March 18. Here are the details (not an affiliate deal).

Next, on to this week’s topic: What’s the real question we need to be asking about artificial intelligence and machine translation (AI/MT)? This topic came to me when I responded to Joachim Lépine’s LinkedIn post about whether AI is a threat to human translators. It’s an excellent question, Joe gave excellent examples, and it sparked a lively discussion! Here’s my take:

What’s different about the current AI/MT boom, when we’ve been leveraging technology in the form of translation memory for a really long time? Everyone knows about it. You know how you get the best information from clients? Being copied in on an e-mail that you weren’t supposed to see. This happened to me a while ago (pre-AI boom) when one of my direct clients accidentally included me on an e-mail saying to a colleague, “I think that the translator has a way to compare last year’s version of the document with this year’s, and then she only translates the new stuff.” This client had no idea what translation memory was; he described it pretty accurately, but the point being that, just as many of us probably don’t know or care about the details of the software used by our accountants, web designers, graphic designers, etc., many translation clients (and here I’m talking about direct clients, obviously agencies knew about TM) and the general public had no clue what translation memory was. Things are really different now, because everyone knows about the new technology; I’ve even had clients (and I’m talking labor unions, not hard-core word people) tell me that they use DeepL for informal translations.
The issue is, most of us are asking the wrong question. We fixate (and I mean, fixate) on AI/MT’s real and perceived weaknesses, of which there are many. It doesn’t get cultural contexts, humor, subtlety, political references, literary references, sarcasm, and the list goes on. I had a client ask me to review a translation in which their MT engine “translated” the name of the Swiss town “Aubonne” as “Augood.” Like, where to even begin? I honestly believe that most professional translators can identify an AI/MT-generated translation in about two sentences.
The problem is, our fixation on the quality of the translation is, in my opinion, the wrong way to go about this. The real question (which I posted in response to Joe’s question) is, how do we find clients who perceive this difference and are willing to pay for it, or who have other reasons for not wanting to use AI/MT, and are willing to pay for those?

I definitely have clients who have moved to AI/MT-generated translations that are then edited by humans. I don’t do machine translation post-editing, not because I have anything against it, but because I find it mind-numbing. And in some cases, I actually don’t think the client made the wrong decision. If you just want to know, “What does this French thing say in English?” DeepL in many cases does a decent job.

But, I have an equal number of clients who ask me, “You don’t use AI-generated translations, do you??” They ask this for various reasons:

They’re afraid, probably correctly, that confidential documents uploaded to an online AI/MT system may end up on the AI provider’s servers. I know that the paid versions of most MT engines promise not to do this, but I still think that it freaks clients out, particularly in certain sectors.I translate for a couple of international family law firms, and here we’re talking about evidence documents that contain people’s most personal information: the details of why their kids were taken away, transcripts of phone calls between them and their abusive ex-spouse, the details of their investment portfolio and how it might be divided in a divorce, etc. etc. Those clients want as few eyes as possible on their documents.
They, too, are word people. “Clients who care” are the best clients, right?? It’s true that every profession has its thing. I’m consistently surprised by interpreting clients who tell me (not being sassy, just being honest), that they don’t perceive any audio quality difference between someone using a laptop microphone and someone using a headset. Whereas to interpreters, this is a huge thing; the other day, my boothmate actually said to the client, “That’s funny, because I can even tell the brand of headset you’re using from the warm tone of the audio: it’s a Jabra, right?.” (And she was correct!) We can’t expect all clients to care about words the way we care about words, but lots of clients do care.
They’ve been burned in the past. Another “best client” category is clients who want a personal relationship with a translator (whether you use AI/MT or not) because they’ve been burned by a different model. I don’t hate agencies, I actually would love to translate for more agencies if more agencies would pay my rates, and I think agencies fill an important role in the language services market. And it’s also true that clients who want consistency, a single point of contact, someone who’s going to know the right questions to ask and the right advice to give, are better off with a freelancer. As an example, I was translating some materials for one of my French university clients (they have “international programs” that are taught in English, into which they accept students who speak zero French). And I asked them, had they thought about any services or programs for their international students’ parents? Parent relations are pretty much nonexistent in the French university system, and I think they thought I was joking, until I gave them the example of my own daughter’s university, which hosts welcome events for parents, a family weekend, and maintains a parent Facebook group. This is a high-priced private university, and I told them that many American parents funding their kid’s education would be looking for this type of thing. AI doesn’t know this!

There are lots of pros and cons to AI, but I honestly think that, “Is there a difference between AI and human translation?” (yes, of course there is!!) is the wrong question, and that we’re better served by asking, “How do we find clients who perceive this difference and are willing to pay for it?”

Corinne McKay (classes@trainingfortranslators.com) is the founder of Training for Translators, and has been a full-time freelancer since 2002. An ATA-certified French to English translator and Colorado court-certified interpreter, she also holds a Master of Conference Interpreting from Glendon College. For more tips and insights, join the Training for Translators mailing list!

Oct 17 2023
Corinne McKay

MT/AI: Where do I think this is all headed?

These days, so many people are asking “Where is this whole machine translation/artificial intelligence (MT/AI) thing headed? Are we all going to be out of a job in a couple of years?” So I decided that for this month’s master class on direct client marketing, I would just prepare a few slides on this whole topic. People seemed to find it helpful, so I thought I would share a summary with you in this week’s post.

On the one hand…

Our imminent demise has been predicted for over 30 years. I graduated from undergrad in 1993, and one of my professors told me, “I wouldn’t encourage you to go into translation, because I think that computers are going to completely take that over pretty soon.” And, 30 years later, we’re all still here!
I’m 52 years old. As long as my body and mind hold up, I plan on working full-time until I’m 70ish, and I’m not looking for another career. Likewise, if my 21 year-old university student daughter wanted to become a translator or interpreter, I would perhaps give her some caveats (the language industry is a lot different from when I started 20+ years ago), but I would feel very positive about that decision, and I would encourage her to pursue a career in translation or interpreting.
DeepL and similar tools have gotten a lot better, particularly for information-only translations (“What does this say in German??”), and I still feel that any good translator can tell you in about three sentences that a translation was done by a computer. To me, MT has no style, no flair, no creativity, and that is a lot of what my clients pay me to do.
I also think that interpreting is, in general, more future-proof than translation. I do think that automated captioning of recorded materials is getting a lot better, and that in very controlled settings, we could see automated foreign-language captioning making some inroads. But I think that when it comes to “the magic earpiece” that turns a live speech or presentation into another language in seconds, those of us working in the industry today don’t have a lot to fear.

On the other hand…

When people say that this iteration of MT/AI feels different, I agree. This iteration is not so much “eating the market from the bottom up,” as it is making inroads into the part of the market that most translators want to work in: mid-market agencies. It seems to me that most mid-market agencies are moving toward (or already using) a model where almost any file that is computer-readable is going to be translated using MT and then edited by a human.
I think that there are lots of freelance translators who have cruised along for years, if not decades, working for a few mid-market agencies paying maybe 11-15 cents per word, with the translator doing little to no marketing and essentially having a steady job where they set the hours. I think that many of these agencies are now offering mostly editing work, at hourly rates that are a lot lower than what their translators were making per-word for translation. Lots of these mid-career translators have never worked with a direct client and, honestly, have no desire to do so, because they just want to translate.

And where do I think this is headed?

Most importantly, I may be totally wrong. People much smarter than I am get technology predictions wrong all the time. I mentioned in the master class that I recently saw a reference to an article from 2012 (!!), in which Google co-founder Sergey Brin said that he thought self-driving cars would be so ubiquitous by 2017 that human drivers would be in the minority. Annnnnd, that’s not true by a long shot, even 11 years later. But, with that caveat, here’s what I think we may be looking at.
I think that translation may follow a similar trajectory to what happened to web designers between about 2000 and 2015. I had a website in 2002. At that time, the only way to get a website was to hire a web designer who hand-coded the HTML for you in a fairly artisanal manner. Then came the earliest versions of WordPress (2003) and Squarespace (2004), and pretty soon, most people who wanted a basic website didn’t hire a designer at all; they used a website creation tool.
However, it’s not like web designers disappeared as a species. And in fact, I would hazard a guess that the work done by most web designers in 2023 is both more interesting and more lucrative that the work they were doing in 2002. I honestly think that if the cream rises to the top of our industry, so to speak, it’s the most interesting and lucrative work that will continue to be sent to human translators.
I also think that in the future, it will be difficult, if not impossible, to earn a healthy income (let’s say 90-110K or higher) without working for direct clients, and most importantly, without having good business skills. I think the world of “like a regular job where you set the hours,” where you’re going five years without doing any marketing, is going, if not gone, and we’ll probably see a fair number of translators drop out of the market or change careers because of that factor alone.

I hope you find these unscientific predictions (!!) helpful. If there’s anything you want to tell me, just put your reply in the comments!

Jan 29 2009
Corinne McKay

Thoughts on machine translation

There’s been a lot of buzz lately about machine translation, with translators and translation industry watchers either condemning machine translation as useless and damaging to the profession or touting it as the next big thing in the industry. ATA President Jiri Stejskal wrote his most recent column for the ATA Chronicle about MT, reader Polly-Vous Français sent this link to a Financial Times blog about machine translation’s role in the blogosphere, and last evening the Colorado Translators Association sponsored an excellent presentation by Cris Silva of All in Portuguese, which focused on integrating machine translation into your existing suite of computer-assisted translation tools.

Cris’ presentation, which she also gave at the recent ATA conference with Giovanna Boselli as a co-presenter, was interesting in that she used the ATA certification exam grading scale to score the output from Google Translate. Out of three sample texts that Cris submitted to Google Translate, two would have failed the ATA certification exam and one would have passed (for what it’s worth, I think that 30% is a better passing rate than what is achieved by the humans who take the ATA exam!). So, what’s the outlook for MT? Friend? Foe? Colleague?

I’ll go on the record as saying that I think that most translators are much too paranoid about machine translation. MT technology has come a long way, and I think that we’re on the cusp of its being considered a standard productivity tool for translators, much as translation memory is today. However, I think it’s not worth losing sleep about machine translation sending human translators the way of the telegraph operator. For an industry leader’s outlook, check out 5 Questions with Renato Beninatto on Sarah Dillon’s blog “There’s Something About Translation.” In this interview, Renato makes the case that the translation industry is in a bind, with “good translators…scarce and becoming scarcer” and the demand for translation growing at 15-20% a year. Companies are turning to MT as an option because their volume of content demands it, not because they want to avoid human involvement in the process.

So, I decided to do my own unscientific test of what Google Translate is producing these days, using three different texts that are similar to what I might be translating on any given day: fluffy, dense and flowery. Let’s see how it did.

For the “fluffy” test, I pulled a tidbit from the French Wikipedia page about France’s first lady, Carla Bruni-Sarkozy. Let’s read about her life pre-Nicolas!

Alors qu’elle vit avec l’éditeur littéraire Jean-Paul Enthoven, elle entame une liaison avec le fils de celui-ci, Raphaël Enthoven, qui était alors marié avec Justine Lévy. En 2001, elle a un fils, Aurélien, avec Raphaël Enthoven[19]. En 2004, elle est l’un des personnages du premier best-seller de Justine Lévy Rien de grave. L’auteur, fille de Bernard-Henri Lévy (dont l’éditeur historique et meilleur ami n’est autre que le père de Raphaël Enthoven), y expose son important passage à vide et sa période de reconstruction à la suite de son divorce avec Raphaël Enthoven, parti avec Carla Bruni.

And the result from Google Translate:

While she lives with the literary editor Jean-Paul Enthoven, she began an affair with the son of the latter, Raphaël Enthoven, who was then married to Justine Lévy. In 2001, she has a son, Aurélien, with Raphael Enthoven [19]. In 2004, she is one of the first bestseller by Justine Lévy Nothing serious. The author, daughter of Bernard-Henri Lévy (including historical editor and best friend is none other than the father of Raphael Enthoven), sets out its important transition to vacuum and its period of reconstruction following his divorce from Raphaël Enthoven, who with Carla Bruni.

At first, with the exception of its confusion over the use of the present tense in French (which sometimes drives me crazy too!), things are looking good. Then, Mme. Bruni-Sarkozy herself becomes a bestseller, rather than a character in one. Finally, the “transition to vacuum” following “his” divorce is a complete disaster. This passage highlights some of the areas in which MT really struggles, for example confusion over the fact that French does not have separate words for “her” and “his.” The key question: is this editable, or could I have translated it from scratch in less time? I’d pick to redo this one entirely.

Next, let’s move on to something dense. Here’s an excerpt from the Quebec department of labor’s information about its purpose:

Emploi-Québec agit à la fois comme producteur, utilisateur et diffuseur d’information nationale et régionale sur le marché du travail.

Son rôle de producteur de données lui permet d’offrir une information de qualité, fiable et à jour.
Utilisant ces données dans sa prestation de services, Emploi-Québec occupe une position privilégiée pour bien cerner les besoins d’information des personnes et des entreprises.

Google Translate tells us that:

Emploi-Québec acts both as a producer, user and disseminator of information on national and regional labor market.

* His role as a producer of data allows it to provide quality information, reliable and current.
* Using these data in its services, Emploi-Québec occupies a privileged position to identify the information needs of individuals and businesses.

Here, I’d say that with the exception of some misplaced modifiers and the pesky problem of French referring to all nouns as “he” or “she,” we’re moving closer to something that’s editable, more along the lines of what one would expect from a non-expert human translator.

Now how about something flowery. Here’s a passage from a project I worked on last year, a promotional document for an arts festival in Senegal.

Faut-il s’étonner que, portés par cette lame de fond, les responsables africains aient, à leur tour, pris la décision historique de transformer l’Organisation de l’Unité Africaine (O.U.A.) en Union Africaine (U.A.) ? En reprenant à son compte la vision prospective de Kwame Nkrumah, la nouvelle Union Africaine a donné à l’Afrique de nouvelles ambitions.

And the result:

Is it any wonder that, carried by the slide background, African leaders have, in turn, took the historic decision to transform the Organization of African Unity (OAU) African Union (AU)? Showing in his account of the vision of Kwame Nkrumah, the new African Union has given to Africa for new ambitions. Complete their destiny.

Wow. To quote that immortal line from Cool Hand Luke, “What we’ve got here is a failure to communicate.” Even someone who doesn’t speak French can appreciate the exuberant incomprehensibility of this passage. Slide background? His account of the vision? And the OAU and the AU are now the same thing? We’ll just let Google Translate complete its destiny on passages such as this.

So, my unscientific observation would be that for straightforward texts, this type of MT is almost at the point where it could be a useful add-on to a translator’s tool suite. Many translation environment tools now incorporate some time of MT module, and on a text such as the second example, I would probably be willing to try this feature. For more nuanced texts, I think that the human touch is still critical