Things are moving fast in translation, and the fallacy mongers are trying to keep up.
- AI will never replace human translators.
Professional human translators were being replaced by MT long before the AI hype and AI business feeding frenzy took hold. With already implemented improvements, AI translation engines are now capable of replacing most human professional translators, given that a sufficient supply of willing post-editors is available, and that has proven to be the case.
- AI isn’t putting you out of work, people who use AI are. AI is just a tool, so translators shouldn’t feel threatened AI, but should rather use AI as a tool.
This commonly heard AI toolist mantra totally ignores the structure of the translation business as engaged in (or, more significantly, not engaged in) by freelancers. If you are on tier two (where most freelancers are) and can only sell your translations to tier-one entities (agencies), when those entities switch to AI as they have been doing to replace human professionals, you are out of luck, unless you can migrate to tier one and compete with them for clients. That is a feat possible by only a tiny number of freelancers, for a number of reasons that many freelancers realize but are reluctant to come to terms with.
- AI cannot understand cultural nuances, so humans will always be necessary.
It doesn’t much matter. The overwhelming majority of translations that are paid for and have supported the great majority of translators have nothing to do with cultural nuances and are not affected by culture. Those translations might require expertise in a specific domain, but culture very rarely enters into the picture for translators doing this mainstream translation work. Not everything is marketing or cultural; in fact, very little is.
- Translators should just refuse to do post-editing of AI output.
I would not place hope on this happening. Too many people have too few options and too many bills to pay. Additionally, the globally connected translation industry has created a distributed labor supply that is uncontrollable and unorganizable, rendering futile any attempt at organized action. It’s not going to happen.
- Translators should form unions and fight back.
Nonsense, for much the same reason. Translators are not employees. This is a feel-good non-starter.
- The AI bubble will burst.
It might very well burst, but it matters not at all to freelance translators, unless they are investing in AI businesses.
Translation-brokering agencies are not going to stop using MT/AI to replace human professionals just because the dreams of fabulous wealth and influence embraced by tech companies are smashed. The agencies will continue replacing human professionals by using AI.
- The supply of translators willing to do dirt-cheap post-editing will dry up, so “then what will agencies do?”
Maybe in a decade or two or perhaps longer. Why? Again, because of the two-tiered, brokered structure of the translation industry and demographics of the freelance translator population. Most freelance translators cannot move from tier two, from which they sell translations to agencies, to tier one, where they would sell translations to direct clients, which are later adopters of AI and care more about quality and connection with a sentient translation provider.
Many of these precarianized translators, faced with not many options—particularly in the case of language or translation graduates with no other expertise—will need to hang on, even getting only low-paid post-editing work.
That labor supply will serve the needs of translation agencies for 10 or 20 years. So, if you’re waiting for things to “renormalize” to where humans are valued (and paid for) again—something I seriously doubt will happen—you need to figure out whether you have enough money stashed away to feed you and your family for 10 years or more when added to the peanuts you will be paid for post-editing. Very few freelancers have that reserve, so “then what will translators do?” I think the answer is clear; they’ll log on to a hamster-wheel agency platform and post-edit AI output. Many are doing it already.
- Horrible consequences will be suffered by companies relying on AI.
This presumes that companies with important translations are going to rely on AI alone. They are not doing that, and need not do that, because of the ready labor supply provided by out-of-work translators to act as AI output repair workers.
That said, how many actual cases can you identify (name names and specific instances, no fair presenting non-verifiable anecdotes) where such dire consequences have become a reality? All we hear are fairy tales, generalities, and hopeful predictions of translation-caused calamities.
There are more fallacies floating around and promoted by people feeling the heat of the AI wildfire at their doorstep. That wildfire will not go out, and promoting aspirational and defensive fallacies won’t help. It’s here and what will happen, will happen. In fact, an awful lot of it has already happened, as we can see colleagues who can do it leave translation for other career paths.
