Habit. And that was just a demo. Plus I knew quoting might be an issue. Perhaps I would change to Finnish if it would be supported here. Here can be found posts by me where I’ve used AI to translate my text. Here is a lot of posts where I’ve used proofreading, and that is sort of AI-translating too, and absolut no one has had any issues, not practical or theoretical. For me that sounds a lot of like double standard or the main reason is totally different than quality of traslations.
But sure. I will use that option in the future, in some scale. But I promote that option heavily on my forum, because it removes, partly anyway [1], that huge language barricade.
Everyone here, and at every global forum, has used to read broken English. And I’ve lived outside Finland several years and learnt not to care how much I violate language. In that context I’ve learnt too that communication doesn’t break because of my vague language skills or when native ones must start simplify theirs use of language. The communication breaks only when there isn’t common language at all.
Finnish is a problematic language for translators, because spoken one (we have a lot local dialects) can be very different than official book Finnish, and almost every one is writing at least in spoken form is some level at casual environments — that’s why translation every now and then don’t do its job, because it just doesn’t know some word or expression. ↩︎
Yes, you’ve read it from your own partial perspective, losing much of the meaning by not caring nor knowing where it came from. Imagine if it were translated…
I guess this is an important point. Is there an indicator of the original language, with an option to revert to it, or to compare the original and the translation? I think that would be critical. Sometimes one reads something coming from a foreigner, and doesn’t know whether what they say is a mistake or a subtlety. Assume goodwill works well with humans, but not AIs, as the latter do not have such property.
It’s great that you never experienced any issues. I did earlier this month. My choice of words was perceived as rude. It was a sentence you can say that way in German without being accusatory in any way. The English translation looks similar, but has a different nuance in meaning. I missed that.
I also tried what the AI translator would have done if I had written my post in German. It made the same mistake I did. The difference would have been it wasn’t me, but AI. I wondered if that would have changed anything. But I am not sure this would have been better. If you feel offended by the choice of words used by a non-native speaker, you will probably feel just as or even more offended by the translated version.
I think the difference in how you say things in different cultures (e.g. high/low context communication) is something AI doesn’t take into account during translation.
And that is changing somehow when non-fluent speaker writes broken English? The reason why you felt you need translation in the first place here is the main point here.
We finns don’t use small talk. At all. From american point of view none from Nordic can small talk (even finns mostly think swedish talk too much nonsense). Germans aren’t known active small talkers either. Americans read that as rude behaviour quite often.
That behaviour doesn’t change if translation is used or not.
Same thing with synonyms and choises what an AI does. It is still better than what previous translators did, or still do. And we are back in the same: which one is bigger threat:
an individual be offended because of style or used word (because she/he/it thinks theirs way is the best and the only way)
using bad english (or what ever now is used as main language), we have here at Meta at least weekly posts, that are harder undestand than my reasonings, because used English is so sub-something
there is no option available at all — that is: turning clock back one week and allow only default ones (really, there is no uniform native English word wide. Or German…)
So we non-native speakers should start worry how native speakers feel some hidden nuances in their cultural context, and everything just because those native ones can’t any other languages Sorry to say this at semi-loud, but that sensitivity limits to two, maybe three, languages Everyone else gives f*** (and one cultural difference more — no matter of word itself, but up here that playing with * is counted really childish, because everyone knows the word, so use the right one, or choose different one, but not that style — and that is something that AI can and will fix )
Well. I don’t have any more points, and this is spinning aroung a circle now. I think this widers conversations and opportunities more than anything before. And if the price tag of that tool is few bad translations and upsettings from someone who is basically the perpetually offended [1], I’m totally willing to pay that price.
So I’m off.
AI-translation, because the original one, sometimes translated The Grump, is really hard to translate ↩︎
I’ve given these posts a new dedicated space in this topic instead of our announcement topic. Most points have been addressed by various other posts in this topic, and I will cover feature related points which have been missed out.
Just a quick note here – Google already translates our posts even before this feature was implemented. When a related query is searched for and unsupported, Google puts the forum through their proxy, e.g.
https://meta-discourse-org.translate.goog/t/content-localization-and-automatic-translations-for-your-community/370000?_x_tr_sl=en&_x_tr_tl=id&_x_tr_hl=id&_x_tr_pto=tc
(yes, that’s a legit search result on Indonesian Google. notice the domain)
However that’s very poor search experience and traffic through a proxy will not give you good data. This is one of the many reasons we’re building this feature.
I’ve spun out new topics from what I’ve gleaned here and I appreciate all the feedback
There are some little quirks with regards to post language detection that I am aware of as seen in post 7 where it was detected as Czech instead of English, and other topics on meta.
Lastly folks, I’m going to request that we take time to think about what we’ve written here and how we’re all on the same page that we want the best for our communities. Wanting to understand each other’s culture is impressive, and I appreciate those who are trying.
There are a few things I am looking out for here (a new topic please)
Is something consistently translated wrongly on meta?
Are folks misunderstanding one another and is anyone uncomfortable?
Is our interface clear over whether content is shown in the original language or translated?
Ultimately Discourse is a construction kit, so we would like to make sure we provide enough tools to set your multilingual community up for success.
Sorry to say this at semi-loud, but that sensitivity limits to two, maybe three, languages 
)
