this post was submitted on 29 Feb 2024
21 points (100.0% liked)

Anime

1809 readers
78 users here now

This community is the place to discuss and ask questions about anime, anime news, and related topics.

Currently airing show discussion threads are created by our resident bot, [email protected]. If it doesn't make a thread for an episode that you want to discuss, see the user guide on the wiki for instructions on how to request that rikka make a thread for you to use.

Check out our wiki to find:

Rules

Related General Communities

rikka

founded 1 year ago
MODERATORS
 

With the recent discussion around AI translation of anime (relevant post), it reminded me of this article/interview I read a little while back in which the topic was brought up with actual, professional translators.

If the topic interests you, check out the full article; it's great. It has lots more info and analysis that what I have put below. I simply picked out a quote from each of the professional translators they spoke to. The full article has much more and great analysis by the author as well.

Zack Davisson (professional translator):

I would say that, like it or not, AI is coming. That genie is not going back in the bottle. And it is improving. The days of Google Translate being a joke are gone. Who knows what AI translation will be like ten years from now? Twenty? Something people need to think about. Hating it is not going to make it go away.

Matthias Hirsh (professional translator):

Machine translations (MT) nearly always need to be proofread or edited by a human, so in many cases, you might as well have hired a human translator. Additionally, having to compete with AI and MT devalues our work. Rates are stagnant among translation agencies as it is. More importantly, however, MT is cutting corners and ultimately leaves the end user with a worse outcome than they would have with a qualified human translator.

Kim Morrissy (professional translator):

Corporations should definitely be more aware of the current limitations of MTL/AI and not see them as a shortcut to reducing labour costs. It’s not just purely a matter of ethics but making people aware that current applications will either see a big drop in quality or require more human labour than they were led to believe. As for the consumers, I don’t necessarily think they should be expected to vote with their wallets purely over this issue, and it would be naive to expect them to do so, honestly. In the future, as AI improves, we may see the debate take on new forms. Personally, I like being optimistic about technological progress. I welcome the tech, just not the way business culture works around it. People have to argue for their rights as workers, otherwise those rights will get eroded over time. That’s just unfortunately how the world works.

you are viewing a single comment's thread
view the rest of the comments
[–] [email protected] 2 points 7 months ago (1 children)

Yeah, it's better, but still far from great. I use DeepL quite a bit to try to make sense of Comic Natalie when keeping an eye out for news that this community would be interested in. However, if I didn't already have a basic knowledge of kanji/kana and trusted the translations blindly, I would be hard pressed to understand half of any news item.

[–] [email protected] 2 points 7 months ago

Yea, and even if the translation is pretty accurate, the actual grammar and sentence structure is awful. An LLM might be able to fix that somewhat, but then you'd have to hope it understands what the ML translation says.