Literotica policy on using AI for translation from non-English languages

[StillStunned]

Please read the long version of what I have written and you'll see I don't disagree with you

I use AI to get different suggestions on parts of a text, and I will eventually (not always) use the one that feel closest to what I had written originally.
I do NOT copy-paste my text to AI just to make a quick translation. It takes me days to make sure that the translation feels acceptable. Whenever I use AI, it's me who decides the words that appear in the translation, not the AI.

The amount of work I put into this makes the final text my own work.

You're still choosing from options provided by the AI. Otherwise, open thesaurus.com and use that.

 

[alexandra_secret]

I feel like when it comes to AI, some people just reject it before thinking about the cases where it can actually be useful...
I don't like AI in general, but it does have some interests here and there.
Once again, as long as it's my decision, it should not matter if the options are coming from AI or from a dictionary. What matters is that what I write in English corresponds to what I had written in French.

Otherwise, I could also tell people to learn French, so they would not have to suffer AI...

 

[AwkwardlySet]

You should write directly in English instead of writing a French version and then translating. Sure, you won't have the same fluency and vocabulary, but that can't be helped. In time, you will get better and improve, and your stories will improve with you. It's the only right way as I see it.

What you are arguing is that because the original story in French is all yours, the AI translation doesn't count as cheating. If only it were that easy. Every language is different and has its own quirks and specifics, so translating a work of fiction from one language into another requires creative work, by definition. Having AI do it for you is still cheating, regardless of the fact that the original story in French is fully your own work.

 

[Bamagan]

I'm sorry but I totally disagree with that. I write in French and I bother to translate to English, but translating is not as fun (to me) as writing and although I can chat in English, I can't produce something satisfying for a written story. Words don't come to me as easily, and my vocabulary is much more limited than in French.
Some other person said that foreign writers "just" have to improve on English, and I also disagree with that. I came here to write stories, not to become an expert in English language.

So yes I've been using AI to help me translate my stories, at the beginning I had some rejections but if you read and correct what the AI gave you, for me it's fine. I use AI to enrich the vocabulary and to give me ideas on expressions to use. When it suggests something I'm not used to, I will dismiss it. That's why I consider that my translation is still mine (it still takes a lot of time, but much less than doing the translation entirely on my own).

Now I totally understand that stories written by AI are not allowed here, even in its own category. But a story that I have successfully published here, it should be allowed that I (and nobody else unless given authorization) publish its translation. Ideally, it should be linked to the original text, eventually with a tag or some indication to let the reader know that AI has been used for the translation.

tl;dr: it's fine to use AI to translate a text, as long as you check and fix the result.

Editing is not as fun as writing for most of us, and using AI as either an editor or a translator is not fine or enriching when you're pretending it's still entirely your own hard work that's being published. When you submitted the story, did you include a note to the editor to tell them that you used an AI to translate? They obviously disagree that it's fine in general, since their policy is not to allow it, therefore they should have the option to make a decision on whether or not it is acceptable in your specific case. I get the impression from your posts that you did not, probably because you know they would likely disagree with you.
You may even be making things even harder for other people trying to publish English translations of their work. It's not hard to imagine the site becoming more critical of submissions from such authors because of suspicion that they're trying to sidestep the guidelines of acceptable content; this could already be the case, considering how many ESL writers have come here to report rejection. The site may even get to the point where it completely rejects basically all translated works in order to reduce their workload, the same way they reject all AI-generated content, even if it's been polished by a human.
I think they would hate to be put in that position (I know I would), but there are potential legal issues related to hosting material that violates, or even may violate, copyright laws where AI is concerned. That doesn't matter to you, most likely, but they probably hope to avoid a scenario where they have lots of AI stuff and need to scour their database to clear it all out in the event of a legal ruling that puts them at risk for hosting it. Something similar has happened to the Illustrated category already, thanks mostly to some different laws, which may indicate how the site would respond to anything which might plausibly have been created with AI.

 

[Five_Inch_Heels]

I feel like when it comes to AI, some people just reject it before thinking about the cases where it can actually be useful...

There aren't any. It's cheating, plain and simple.

 

[Omenainen]

Some other person said that foreign writers "just" have to improve on English, and I also disagree with that. I came here to write stories, not to become an expert in English language.

Not everyone can afford this opinion. There’s only three of us Finnish people, and the other two are not interested in my smut, so if I want anyone to read my stories I need to write them in some other language. So you’re privileged in a way.

I use AI to get different suggestions on parts of a text, and I will eventually (not always) use the one that feel closest to what I had written originally.
I do NOT copy-paste my text to AI just to make a quick translation. It takes me days to make sure that the translation feels acceptable. Whenever I use AI, it's me who decides the words that appear in the translation, not the AI.

If you know the correct words, why don’t you just type them in the first place?

The amount of work I put into this makes the final text my own work.

Actually, no.

 

[TheLobster]

LLMs are great at trawling vast repositories of knowledge in a way that simple keyword-based queries cannot, and producing answers, or at least breadcrumbs to answers, for questions that would require a lot of semi-manual work to resolve otherwise.

In other words, they are great research aides. Actual writing, though? Not so much.

 

[jsmiam]

AI as we know it today, is almost entirely based on “training data”, and particularly in the case of writing and translation, (and all those AI summaries for your searches) that means taking everything the sleazy thieving ai companies could take off of the internet, without permission and presenting it as their own.

In other words, just as there is writing out there, taken without permission, to “train” AI models, there are translated works (original language to new language), also taken without permission, for training purposes. And that’s how and why ai translation tools work. Stolen training data.

In other words, almost everything of value provided by AI today, is rooted in theft. The isssue is the theft. Who cares if somehow the ai tool has some debatable merit to its functionality. It’s stolen. That’s the true issue.

To AI fans: start doing all your shopping at pawn shops. You’ll love the low prices. Sure the things you buy are probably stolen. But you don’t care. Happy shopping!

 

[iwatchus]

In other words, they are great research aides

Except you need to research everything it tells you because a non-trivial amount of it is hallucinated. And many of the hallucinations are plausible, but wrong, so you can't just say "that sounds right"

They are not like an old-fashioned search engine. They do not store all the information, they store tendencies. All they are doing is predicting the likely next word based on the previous n words. N grows as the LLMs get larger, but the basic stochastic parrot nature of them never changes.

 

[TheLobster]

Except you need to research everything it tells you because a non-trivial amount of it is hallucinated. And many of the hallucinations are plausible, but wrong, so you can't just say "that sounds right"

They are not like an old-fashioned search engine. They do not store all the information, they store tendencies. All they are doing is predicting the likely next word based on the previous n words. N grows as the LLMs get larger, but the basic stochastic parrot nature of them never changes.

If you need hard facts, you have to follow the answers to find the original source. But if you don't, since this is fiction we're talking about, then well -- if merely "plausible" was good enough for Dan Brown, it should be good enough for smut

 

[AwkwardMD]

Hey Awk, bar open?

@AwkwardMD

I don't maintain the bar. I just drink here.

 

[Five_Inch_Heels]

Sounds like you need a high school AI bot to do your dirty work.

 

[Meekly_Anna]

Without wanting to offend anyone here - OK, actually with wanting - when I read through the one or other story here, I have the impression that some native English speakers should use AI to translate their stories into proper English.
I, as someone who spent years at the bottom of the English grade spectrum at school, almost ache when I read here. And I don't just mean the spelling, the grammar is also .... remarkable.

 

[Five_Inch_Heels]

Real English ain't what you learn in skul.

People don't talk in structured, styled forms.

 

[onehitwanda]

I don't maintain the bar. I just drink here.

Is there room for me?

 

[Five_Inch_Heels]

Under or on top, dancing?

 

[onehitwanda]

I just did a thing and went off and did something I've been resisting for a long time - which was to do some research on how LLMs work under the hood - since it's becoming of increasing importance in both my real life and this second home.

I can see why they seem almost magical as tools. To someone with fair to middling ability, they feel like they bring a massive leap in the upper bound of what one can accomplish.

Under the layers of gilt and rococo, though, they're just a probability machine that answers the question "given this stream of tokens, what is the most likely next set of tokens"

they're phenomenal achievements. But as an engineer and writer I could never in good conscience trust the unmoderated output of one. And maybe I'm a dinosaur who's destined to go out with a bang (hur hur) but I don't see myself ever willingly using one.

 

[onehitwanda]

Under or on top, dancing?

Just keep the tequila coming and move the other girls back out of harms way.

 

[StillStunned]

Without wanting to offend anyone here - OK, actually with wanting - when I read through the one or other story here, I have the impression that some native English speakers should use AI to translate their stories into proper English.
I, as someone who spent years at the bottom of the English grade spectrum at school, almost ache when I read here. And I don't just mean the spelling, the grammar is also .... remarkable.

All we can do is encourage them to improve. The truth is, though, that some don't want to improve, and some don't see any value in improving. For some people here - writers and readers - it's enough if the words describe an act or a scene. How the words are strung together is completely irrelevant. The gratification, I think, comes from *the idea* of what the story is telling, not the way the story is told.

It's not so far distant from people quoting Monty Python, and everyone laughing because they can picture the scene. It's not the same level of craft as the actual movies and television series, but a few key words will conjure up a scene in people's minds and give much the same effect. And I say this as someone who's spent a quantifiable portion of my life quoting Monty Python.

So some writers feel the need to writer stories with the same level of sophistication as a movie, and others are happy with just providing the key ideas of a scene or two. There's an audience for both.

 

[Meekly_Anna]

All we can do is encourage them to improve. The truth is, though, that some don't want to improve, and some don't see any value in improving. For some people here - writers and readers - it's enough if the words describe an act or a scene. How the words are strung together is completely irrelevant. The gratification, I think, comes from *the idea* of what the story is telling, not the way the story is told.

It's not so far distant from people quoting Monty Python, and everyone laughing because they can picture the scene. It's not the same level of craft as the actual movies and television series, but a few key words will conjure up a scene in people's minds and give much the same effect. And I say this as someone who's spent a quantifiable portion of my life quoting Monty Python.

So some writers feel the need to writer stories with the same level of sophistication as a movie, and others are happy with just providing the key ideas of a scene or two. There's an audience for both.

You are right and you are off topic.

AI is a tool. No more, no less. A very powerful tool, as it appears to us today, but that's all it is.
Perhaps a word about taxonomy. AI is not exactly that. It is not intelligent. It is a programme for recognising patterns and reproducing them. It does not create. It does not ‘think’. Maybe one day there will be programmes that can think, but what we see today is just a pretence.
And so back to the topic: I fully understand that computer-generated stories are not desirable; however, I have no sympathy whatsoever for the fact that translations are undesirable. With this attitude, we should all still be hardening goose feathers in hot sand, cutting them to size with a penknife and then writing our stories on parchment with iron gall ink.

 

[AwkwardMD]

I fully understand that computer-generated stories are not desirable; however, I have no sympathy whatsoever for the fact that translations are undesirable.

They are indistinguishable.

 

[Britva415]

they are great research aides

you need to research everything it tells you

That's actually all right. It's still a good aid because it greatly expedites the discovery of material for you to research in the first place - which is HUGE.

"Needing to research what it tells you" isn't a bug, it's a feature, in the context we're talking about.

Treating it like it's doing the research for you so you don't have to isn't a failure of the tool, it's a failure of the user.

 

[Meekly_Anna]

They are indistinguishable.

Mais, Madame, il ne m'appartient pas de la convaincre de son erreur ; le grand Voltaire ne disait-il pas que nous pouvons et devons tous avoir des opinions différentes ? La liberté d'expression et le droit de la faire connaître par la parole et par l'écrit, c'est l'un des grands acquis de la démocratie. Ce qui était bon et juste pour les Grecs de l'Antiquité l'est toujours pour nous.
Pour paraphraser Evelyn Beatrice Hall : « Je désapprouve ce que vous dites, mais je défendrai jusqu'à la mort votre droit de le dire ».

 

[onehitwanda]

With this attitude, we should all still be hardening goose feathers in hot sand, cutting them to size with a penknife and then writing our stories on parchment with iron gall ink.

Some would argue this would not be the negative that you believe it is.

 

[Meekly_Anna]

Some would argue this would not be the negative that you believe it is.

I cut my first goose feather at the age of twelve and used it to copy manuscripts from the 12th century from a lexicon. I understand the fascination and it gives me great pleasure.
However, I doubt whether this is the right approach to exchanging opinions today.