Will GenAI tools be able to do translation work?

Image by FreePik

Generative AI is largely based on English data inputs and was primarily designed to understand and generate text in English. I’m interested to hear about the experiences our non-English-speaking friends and colleagues are having with GenAI.

When you read data outputs, does it feel misaligned? Sort of foreign?

Or is it bang on? Or not so much? What language are you using?

I know OpenAI has been working on multilingual capabilities to understand and respond in a variety of languages. Apparently, this works to varying degrees of proficiency. These languages include Spanish, French, German, Chinese, Japanese, Korean, Italian, Dutch, Russian, and Portuguese.

How well, or not, do these multilingual capabilities work?

Last year I led ten user research workshops for a fantastic Japanese agency. The cultural differences were profound compared to my primarily Western experience. I used a platform called DeepL to translate my decks into Japanese and then worked with a terrific Japanese interpreter, Fujiko Suda, to ensure the final outputs were sound. I looked at one of my Japanese-translated workshop presentation decks today which inspired me to think about the GenAI capabilities in other languages if OpenAI or similar platforms might be suitable for this kind of translation in the future, and its impact on various cultures, in general.

Would GenAI tools be able to do this type of translation work?

Do you have experience with any GenAI platforms in different languages and or cultures? How is it?

I wrote about this Japanese cross-cultural user research experience if you are interested. The link is in the comments below.