We often hear assertions that translation consumers will be happy with translations done by AI because they are good enough. I think that’s correct. But there’s another good enough that people don’t discuss much.
Translators complain, justifiably, about tumbling rates from agencies and about not having much work other than post-editing of MT/AI output. How much does an agency need to pay translators to do post-editing, the only task that will remain in the near future for former translators?
The answer is clearly an amount that is “good enough.” But good enough for what? Good enough that they can get the former translators to continue to post-edit, and it need not be more than that good-enough amount.
And that is succeeding, for the simple reason that sufficient numbers of former translators have no other options. What they are being paid is good enough to keep them post-editing for agencies.
And here’s a bit of mathematical fallacy that doesn’t get talked about much.
Post-editing is very low-paid work. Let’s say it is paid only one-fifth what a professional translator formerly earned. That’s a figure that can be confirmed from what is being proposed as rates on the click-work platforms so commonly used by freelancers.
Agencies assert that the translators can make up for that rate by multiplying their productivity commensurately. I had someone from an agency tell me just that face-to-face just the other day. While I don’t believe it, let’s give them the benefit of the doubt and say that’s true. What happens then is that the post-editors actually working with a commensurately higher efficiency might earn the same, which would be fine for them, assuming they could stomach doing post-editing. That just means, however, that only 20 percent of the agency-dependent translators will be needed and will be able to survive. That 20 percent happens to be the very outside-chance maximum of translators I think who will still have work by mid-2027. However you approach this, it is clear that most freelancers will see their translation careers ending shortly.
And as this is rapidly playing out, many translators look away or present indefensible arguments for their survival (for example, use AI yourself), and translation organizations like Japan Association of Translators and ATA help them by pretending that freelance translators will just “adapt” and by holding irrelevant events and webinars that don’t address the real problems of translators working in the high-demand fields. That behavior is self-serving and disingenuous, but it somehow goes on without criticism. Comfortable fairy tales continue to be more popular than painful truths.
How long will this insanity continue?