Using plain language for better machine translation in the age of AI

Using plain language for better machine translation in the age of AI

More people today can understand content in other languages thanks to machine and AI translation. These translations can be more understandable if you use plain language principles.

Today’s machine and AI translators cannot match the nuanced and skilled work of professional human translators. But machine translations1 are useful for getting the gist of information. I certainly love Google Translate for confirming signs, instructions, and ingredients lists!

Especially online, where English is dominant, making your text machine/AI-translation-friendly could help more people find, understand, and use your messages and help diminish digital and information gaps.

How to make a text more machine/AI translation-friendly

So how can you make a text more machine/AI-translation friendly? The answer (surprise, surprise!): use plain language.

In this case, your audience is a broad and ambiguous group of “international” readers and a machine that is great at finding patterns but not so good at reading between the lines. Vague and complex writing is more likely to be mistranslated, so you can get better results by using simpler and specific language.

Some techniques you could consider:

Any other tips? Please share in the comments!

Ethical considerations

There are many, many questions surrounding AI translation ethics and quality. As far as quality is concerned, I firmly believe that the machines are nowhere near the level of human professional translators. But we do live in an age where people use machine translation to get the gist of content.

If you don’t read English, you are seriously disadvantaged: a 2023 study by Rest Of World found that more than half of publicly accessible websites use English as the primary language, although “native” English speakers make up just 5% of the global population.

Machine translation thus helps more people find, understand, and make use of information—access that English speakers take for granted.

As for the many other questions (AI’s vast energy and water costs, disregard of intellectual property rights, replication of bias, exploitation of workers)—for now, each of us can only try to stay informed and decide how to use AI. (I believe in the responsible and ethical use of AI and do not use generative AI for editing work.)

If you are looking for an informative source for AI updates that’s not overwhelming, I recommend Mignon Fogarty’s AI Sidequest newsletter.

Summary

  • We are living in an age where our digital content will likely be machine/AI-translated.
  • Make it easier for your global readers by using plain language principles.

I tried using ChatGPT to help copyedit text that a client translated with ChatGPT. Was it useful or useless in working with translations?

Cover image by Mohamed Hassan from Pixabay.

Footnotes

  1. Machine translation and AI translation are not the same thing but I’m lumping them together for this article.
  2. Machiko Asai demonstrated this in her 2020 book, 伝わる短い英語―アメリカ、イギリス、カナダ、オーストラリア 政府公認 新しい世界基準Plain English (Rakuten affiliate link).

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.