The Slow Death of Human Translation

A short piece on the direction of the industry with some tips for the future

The only ones who argue that localization brings in revenue are the localization experts and companies. The only ones who argue that human translators and interpreters are essential are, you guessed it, human translators and interpreters.

Just the other day, my half sister, who had never done anything even remotely related to localization, praised how great the translations generated by AI are. That goes to show that people who know absolutely nothing about translation or localization now think they are experts in the said segment just because it’s all readily available and free and instant.

Or, at the very least, they think they can judge the quality and, therefore, its worth.

And we can’t blame them, and in their own way, they are actually right.

There is no going back in the translation and localization industry. AI has completely redefined the process, quality requirements and especially the perception of the industry in the minds of “translation buyers” – who will slowly, but surely, cease to exist.

Human translation will become obsolete not directly due to its cost or the complicated process which it entails. It will become obsolete due to the perception in the minds of translation buyers – who are, after all, the most important piece of the puzzle.

If you have a service, but no one to sell it to, you might as well close shop.

At this stage, there is no point in battling and arguing against the use of AI in localization. There is no point in arguing that human translators and interpreters won’t be replaced – because they already are being replaced.

There is a couple of possible ways to deal with this situation:

Start an AI startup, which will use what I call the “AI squared model” – where AI translations are checked by AI models, without or with the least possible involvement of human agency. This way, the price of translations should soon approach zero (many soon to be former customers believe that the service of translation had already reached the price of zero).

Move into a different related field, ideally with AI involved, and leverage some of the skills acquired as a linguist and translator. This will probably involve training AI models for various companies – mostly in the form of chatbots or data annotation etc.

The thing is, you can always keep translating for fun or for pure enjoyment. I am planning on translating a book on the beginnings of terrorism in 19th-century Ireland. But that’s not something I am going to make a living off of.

Translation as a form of “making a living” had ceased to exist. The important thing to do now is to adapt. After all, large language models still work with language in some capacity (albeit only statistical), and there still seems to be a need for human agency in their use. Translators can also leverage their knowledge of multiple languages in multinational corporations and companies, international development, development of AI chatbots in multiple languages, data annotation etc.

You might even become a lead of AI development – as it seems no one really knows what the current AI is all about; what the companies do know is that they need to get on the bandwagon. And quickly.

I believe the sooner you start adapting to the changes in the industry, and to the changes in the perception of the soon-to-be-former clients (former translation buyers), the better off you will be in the long run.

Last but not least, anyone enrolled in a university translation program or anyone who intends to become a translation major – don’t. Unless you throw in machine learning/AI or any other skills that could be monetized.