I was listening to an interview with a senior EU translator several years back, and he said that these days, he normally does the first pass with Google Translate, then manually cleans things up. My guess is that to some extent, most human translations likely incorporate some AI translation already.
My SO did translation as a contractor for a little while, and that’s what they did too. Run it through a translator, and fix whatever it messes up. A lot of the output is totally fine, but not all of it, so you need someone experienced with both languages to make sure the result is good.
Correct. But the AI bro here think AI translation is the final work, while translator that use google translate still required the language knowledge to proofread.
Pure machine translation would indeed be sloppy, but games have (unfortunately) done it before. An automated 1st pass with a last check from a human contractor seems reasonable for a studio about to fold.
AI Bro is pretty specific. To me, its evangelists worshipping nebulous ideas and figures like Altman or maybe Musk, looking down on others for not “understanding” how amazing their vision of AI is, all in on the enshittification and impracticality, all in on the raging hype.
It feels very much like crypto fanaticism.
Even if we interpret OP as cynically as possible (lazy AI-only translation when they have another option)… that’s bad, but not “AI Bro” to me.
If it saves your company from bankruptcy, then why not do it? The developers even said it got a little out of control.
Sure, you can call me an “AI Bro” lol. But I’m just being real. Using AI or finding another job? Use AI all day and make better choices with the next title.
“Save” is a bit stretch, if they officially released a halfassed translation(which AI translation without went through a layer of human emotion is), you will be damn sure the reception is gonna be negative, which would negatively impact the sales and also reputation they have.
There shouldn’t be any problem in using AI to translate something, translation is more or less static. Its no different than someone using a calculator for mathematics equations.
Localizers will still need to check the AI output for contextual accuracy, but they will be able to complete this faster as they can essentially skip a step.
My only issue with translation currently is that localizers often go too far with the liberties they take. Its necessary to ensure people from another culture can understand what is happening. For example, in a language that has no word for “rye bread” or a saying like “you are what you eat” specifically, the localizer may substitute the closest word or phrase that conveys a meaning as close as possible to the original. What is not okay is completely altering large portions of the work because of the localizer’s personal opinion. And unfortunately, because this is entirely on the localizer, no amount of AI can help prevent that. Unless translation AI can be so good that it can even understand context from the various bits of text needing to be translated. Then the developers can just use it themselves. But AI has a while to go before it gets to that point.
a member of my extended family spent a decade or so as an interpreter. when they got their certification, the little ceremony we attended had a panel of level 5 interpreters (the highest score you can get on the test they were taking) interpret a speech from one of the more popular interpreters in the local union. i doubt one of them used the same word at the same time. when we did the national anthem, because of course we did, there were two scripts because two languages. i hate saying it, but it really kind of depends what you’re translating/interpreting, but usually translations can be fluid.
Quality not withstanding you’ve got to pay for access to the model or electricity to run your own local model, pay people to run the lines into the model and stitch them back into the game and pay people who speak the language to proof read the outputs to ensure it’s not giving you gibberish.
And if you’ve got voice lines now that’s a whole other can of worms of paying for TTS ai models, paying for audio mixing specialists, inserting the lines into the game, paying to once again have a speaker of the language QA test the output.
The electric costs aren’t nearly as high as people think. For huge datacenters, yes, but that’s because they’re processing requests for hundreds or thousands of users simultaneously. For a studio using it to translate lines for a single game, they could easily get away with doing it locally, and effectively for free. You can train your own local model on a consumer-grade PC without any issue, and it’ll still run just as fast as the big server farm-powered models.
My roommate has been playing with a bunch of different local AI models on his own PC for a couple years now. There’s been no discernable change to our electric bill. His PC draws more power playing an anime waifu gacha game than it does training/generating AI.
Translate using AI and call it a day.
Yes, this will turn their potential studio shutdown into an guaranteed studio shutdown! Problem solved.
I was listening to an interview with a senior EU translator several years back, and he said that these days, he normally does the first pass with Google Translate, then manually cleans things up. My guess is that to some extent, most human translations likely incorporate some AI translation already.
My SO did translation as a contractor for a little while, and that’s what they did too. Run it through a translator, and fix whatever it messes up. A lot of the output is totally fine, but not all of it, so you need someone experienced with both languages to make sure the result is good.
Correct. But the AI bro here think AI translation is the final work, while translator that use google translate still required the language knowledge to proofread.
I don’t think OP came off as “AI Bro.”
Pure machine translation would indeed be sloppy, but games have (unfortunately) done it before. An automated 1st pass with a last check from a human contractor seems reasonable for a studio about to fold.
“and call it a day” is all the sign i need.
Not following that at all…
AI Bro is pretty specific. To me, its evangelists worshipping nebulous ideas and figures like Altman or maybe Musk, looking down on others for not “understanding” how amazing their vision of AI is, all in on the enshittification and impracticality, all in on the raging hype.
It feels very much like crypto fanaticism.
Even if we interpret OP as cynically as possible (lazy AI-only translation when they have another option)… that’s bad, but not “AI Bro” to me.
If it saves your company from bankruptcy, then why not do it? The developers even said it got a little out of control.
Sure, you can call me an “AI Bro” lol. But I’m just being real. Using AI or finding another job? Use AI all day and make better choices with the next title.
Yeah honestly I agree with you.
But like others said, not sure iffy translations would be enough to save the company.
“Save” is a bit stretch, if they officially released a halfassed translation(which AI translation without went through a layer of human emotion is), you will be damn sure the reception is gonna be negative, which would negatively impact the sales and also reputation they have.
deleted by creator
There shouldn’t be any problem in using AI to translate something, translation is more or less static. Its no different than someone using a calculator for mathematics equations.
Localizers will still need to check the AI output for contextual accuracy, but they will be able to complete this faster as they can essentially skip a step.
My only issue with translation currently is that localizers often go too far with the liberties they take. Its necessary to ensure people from another culture can understand what is happening. For example, in a language that has no word for “rye bread” or a saying like “you are what you eat” specifically, the localizer may substitute the closest word or phrase that conveys a meaning as close as possible to the original. What is not okay is completely altering large portions of the work because of the localizer’s personal opinion. And unfortunately, because this is entirely on the localizer, no amount of AI can help prevent that. Unless translation AI can be so good that it can even understand context from the various bits of text needing to be translated. Then the developers can just use it themselves. But AI has a while to go before it gets to that point.
a member of my extended family spent a decade or so as an interpreter. when they got their certification, the little ceremony we attended had a panel of level 5 interpreters (the highest score you can get on the test they were taking) interpret a speech from one of the more popular interpreters in the local union. i doubt one of them used the same word at the same time. when we did the national anthem, because of course we did, there were two scripts because two languages. i hate saying it, but it really kind of depends what you’re translating/interpreting, but usually translations can be fluid.
That still has associated costs my guy.
Quality not withstanding you’ve got to pay for access to the model or electricity to run your own local model, pay people to run the lines into the model and stitch them back into the game and pay people who speak the language to proof read the outputs to ensure it’s not giving you gibberish.
And if you’ve got voice lines now that’s a whole other can of worms of paying for TTS ai models, paying for audio mixing specialists, inserting the lines into the game, paying to once again have a speaker of the language QA test the output.
The electric costs aren’t nearly as high as people think. For huge datacenters, yes, but that’s because they’re processing requests for hundreds or thousands of users simultaneously. For a studio using it to translate lines for a single game, they could easily get away with doing it locally, and effectively for free. You can train your own local model on a consumer-grade PC without any issue, and it’ll still run just as fast as the big server farm-powered models.
My roommate has been playing with a bunch of different local AI models on his own PC for a couple years now. There’s been no discernable change to our electric bill. His PC draws more power playing an anime waifu gacha game than it does training/generating AI.
Your ideas are bad and you should feel bad