The same is happening to a lot of website owners. You lose half or more of your traffic to AI summaries trained on your own content. The cost of producing original content is the same as before.
It makes research harder too, since more and more public information is infected by AI content. Both published posts and internet discussions are tainted.
And then the AI companies threaten to crash the whole economy if we don't let them do it.
For the millionth time, piracy isn't theft. It's copyright violation, not copyright theft.
No, a reduction in traffic is not sufficient to conclude that a copyright violation has occurred. Sure, it might have. Alternatively it might have produced a lossy summary in which case the reduction in traffic raises some difficult questions about the value of the original work.
In other cases an LLM can synthesize a genuinely useful explanation of a subject that is precisely tailored to the needs of the asker. In those cases the machine output might well prove more useful to the asker than any single original reference would have.
For something like news where what you're paying for is timely delivery it makes sense to restrict automated (not just LLM) access for the first few days because a similarly timely summary will capture the majority of the value proposition of your service.
That's not typical though. For example, I'm certainly not going to be satisfied with a summary of the plot of a book I'm interested in. Would you want to watch a 10 minute highlights reel in place of a 2 hour feature length film?
For the millionth time, the reason we have copyright in the first place is to encourage creation of original creative works. This is clearly stated in the US constitution (and similar phrasing is found in the relevant legal texts of other jurisdictions).
You can apply obsolete legal tests that have been used to enforce this principle all day long, but the central question remains: Does generative AI encourage creation of original creative works?
If the answer is "no", which it clearly is, then whatever laws and legal tests exist to enforce IP rights need to be amended - or the constitution does.
The problem with your definition is that the pre-copyright history is a good example of why we don't need copyright. The US applied copyright laws very late (similar to what China did recently), which led it to be the nation where the citizens read the most in the 19th century. This then led to the cultural explosion we know.
Free reproduction of "original creative works" fuels original creation, too, while creating tight monopolies over intellectual works and universes has led to a decreased creativity around them.
See the dire state of the US film making industry, as an example. Or the vast amount of bizarre lawsuits such as the one for the "Bittersweet Symphony".
I believe that no, copyright does not encourage the creation of creative works, because the mechanism of an exclusive economic monopoly on a creative work supresses more expression than it encourages. Copyright may have hade more validity when the printing press was new, but in a modern context is the wrong mechanism. Instead of encouraging creativity we have instead encouraged capital acquisition and management.
Surely the issue you speak of is largely due to duration? What's wrong with an author (for example) having exclusive rights to a book for 20 or 30 years? Shouldn't that be expected to increase his ability to create additional works?
So if we're talking about the extraordinary power of the state to enforce a right, it has to be in the interest of the public. What is the desirable state we're getting at? Are we maximizing the published creative works, are we concerned with making writing a profession, do we even want culture to be tied to economic activity? Seeing what fans produce with fiction and costumes; I have no doubt that we would have a vibrant and active world of art and fiction without an economic incentive. Do we as a society value the market or the art, and to what degree each?
I agree, but I think there's a second relevant question as well: Does generative AI output original creative works of its own? Obviously the goal here should be maximizing societal benefit. Specifically encouraging the creation of works that we (as a group) find directly useful or otherwise desirable. At least to my mind, human exceptionalism is an explicit non-goal.
I'm already finding the ability of LLMs to synthesize useful descriptions across disparate sources of raw data to be immensely useful. If that puts (for example) scientific textbook authors out of a job I'm not at all sure that would prove to be a detriment to society on the whole. I'm fairly certain that LLMs are already doing better at meeting the needs of the reader than most of the predatory electronic textbook models I was exposed to in university.
> If the answer is "no", which it clearly is,
Why are you so certain of this? It clearly breaks many (most?) of the existing revenue models to at least some extent. But we don't care about the existing revenue models per se. What we care about is long term sustainable creation across society as a whole. So are consumer needs being met in a sustainable manner? Clearly generative AI is (ever increasingly) capable of the former; it's the latter that requires examination.
Normalizing the idea that any such thing as "copyright theft" actually exists would certainly be transformative - and what a stifling mess that would be!
I'm not too fussy about the choice of words. I do the work, they use their monopoly to capture the value. There is no way to opt out, no permission asked. I had a bitter chuckle when Microsoft said that they might "lose social permission" if they can't find good uses for AI.
> It makes research harder too, since more and more public information is infected by AI content.
God, how I hate the billions and billions (sorry Carl Sagan) of pages that look like they're going to have information, but just repeat your question and expand it to 20 or more paragraphs.
"You want to know about how much torque to tighten wheel nuts for a Ford Focus. Torque to tighten wheel nuts is an important aspect in securing the wheels after replacing them...". God damn, it's Eliza being abused to serve the attention economy, and you notice this and wonder if the page would have the correct information or it would just be another bucket of shit pretending to be information... So you think "Let's skip all of that and ask AI...".
Someone, somewhere has to put some information online for the first time. I am one of those people.
This sort of thing was a problem before too. I had a few copycats myself. The difference is that those copycats competed on the same level, and this sort of shallow copying was still too expensive. Google can automate plagiarism at a scale and just shut the competition down.
They are the modern Terminal Railroad Association.
Would like to point out that professional translation has been under pressure for much longer than AI.
I have friends that made a descent buck 20-30 years ago translating technical documents like car manuals. Over the years, prices fell from quarters per words to fractions of a cent.
And even though machine translation was barely existent, tools were used to argue higher productivity and therefore lower prices.
This. Google, when it went to a more neural-based translation system about a decade ago, along with DeepL put a serious dent in translation work - particularly around gigs that don't require extremely rigorous accuracy like legal or medical.
Several of my friends who used to work full-time as translators are now supplementing their income with side jobs like foreign language teaching, proofreading, and similar work.
Recently I've noticed that the subtitles on TV shows have become garbage. Often the translations makes no sense in the context of what is happening, but are literal translations of what is being said on screen. I assume it's some kind of AI translation, but who knows. Maybe the translator is just stupid.
To be clear, we had bad subtitles before as well, but that was due to the translators lack of cultural understanding. Most of it made sense at least. Now it's just straight up bad and often makes no sense. Sad!
> While using AI tools for everyday tasks like finding directions is “low-risk,” human translators will likely need to be involved for the foreseeable future in diplomatic, legal, financial and medical contexts where the risks are “humungous,” according to Benzo
Now it's a classic, you need an expert in order to check the work of the machine, because the "customer" is by definition not able to do it.
Aside from highly technical domain, in purely literary works, I think that the translator is a co-author - maybe IP laws acknowledges that already? I remember the translation of E.A. Poe by C. Baudelaire for instance; I think you could feel Baudelaire's style because it is a lot "warmer" than Poe's. I've also read a translation of a Japanese novel and I was quite disappointed with it. I don't know Japanese but I have read/watched quite a few mangas/animes, so I could sense the speech patterns behind the translations and sometimes thought they could have made better choices.
In any case, one will still need a translator who is good at "prompt engineering" to get a quality translation. I don't know. Maybe translators can add this skill to their CV, so they can propose quick-and-dirty/cheap translations, or no-AI high quality translations.
Some suggest "no-AI" labels on cultural products already - I think if it becomes a reality it will probably act as "quality signaling", because it is becoming more difficult every year to tell the difference between AI and human productions. It won't matter if what you read was written by an AI or a human (if it quacks and looks like a duck...), but what the customer will probably want is to avoid poorly-prompted machine translation.
> it will probably act as "quality signaling", because it is becoming more difficult every year to tell the difference
Note that this only applies to something like a translation where there's some notion of a "correct answer". For other cultural products it's irrelevant (as you say, if it quacks like a duck ...).
Quality signaling is really only necessary in situations where an upfront investment is required and any deception is only revealed sometime later upon use. Safety critical systems such as airbags are a model example of this - a counterfeit of deficient functionality won't be discovered until it deploys, which in most cases will never happen.
That said, while I certainly can't speak to business or diplomatic translations, when it comes to cultural works (ie entertainment) the appeal of machine translation to me has been gradually increasing over time as it gets better. I don't generally find localization desirable and in some cases it even leads to significant confusion when a change somehow munges important details or references. Confusion which I'm generally able to trivially resolve by referencing machine output.
AI translations are getting good. I've been working on our company website in the last month. It's a static website. So, I use Agentic coding tools to implement both content and technical changes. A simple prompt "align the translations with the English version for all pages" and some guard rails in the form of skills and AGENTS.md instructions are doing most of the heavy lifting. To be clear.
These translations are not perfect, yet. But good enough for my needs. Any professional translator services would in any case be beyond our budget. The advantage of using agentic coding tools here (enabled by using a site generator rather than a CMS) is that I can get systematic about dealing with jargon, SEO, and frequently used phrases. I simply document all that and instruct the tool to refer to that. The funny thing is that most of the models are pretty good at fixing their own mistakes if you just ask them too. I asked it to look for examples of "denglish" (German English) in its own German translations and then to fix it. It found a few examples and the suggested fixes were fine.
A lot of people are focusing on the negative here. I like to look at the positive. We're approaching the moment where any person on this planet will be able to communicate directly with any other person on this planet without the need for translators. The tools already exist for this. But they need a lot of work on quality.
A second point here is that the role of English as the most popular intermediary language is disappearing as well. I'm not a native speaker. When I talk to foreigners from wherever, it's mostly in (bad) English. By definition that limits me to talking to people that have had enough education and exposure to English. This is very limiting. A lot of the people we need to talk to here in Germany aren't all that comfortable speaking English.
And where would you put "billions of people now have access to better translations on demand"?
People talk about business as though only the owners of the business benefit. Everybody else pays the price. But aren't the main beneficiaries all the people using these services?
> billions of people now have access to better translations on deman
As a German speaker, I experience the quality of German language technical documentation steadily declining. 30 years ago, German documentation was usually top notch. With the first machine translations, quality went notably down. Now, with LLM translation, it's often garbage with phrases of obvious nonsense in it.
This is especially true with large companies like IBM, Microsoft or Oracle.
I guess the situation is better for languages where translations only became available with LLM.
"Meanwhile, the poor Babel fish, by effectively removing all barriers to communication between different races and cultures, has caused more and bloodier wars than anything else in the history of creation."
Where does it say that? I have never read any Book of Genesis that says "God destroyed the tower".
Also it doesn't say "God was afraid". God doesn't have negative emotions like that. God plans out everything, so He is not "afraid" in the human sense.
In fact, I am fairly certain that the mythical "Tower" for the Jews was sort of a parody of the Pyramids of Egypt and the Ziggurats of Mesopotamia. They were essentially mocking their ancient neighbors in the Levant for such a frivolous project that they believed really didn't honor God, but increased their arrogance and hubris.
In fact, the Sumerians worshipped a god named "Sin" https://en.wikipedia.org/wiki/Sin_(mythology) and it is believed that the "plain of Shinar" and the "wilderness of Sin" are cognate with this term, and therefore represents the ancient deity that was worshipped in that particular case.
For Egypt, the pyramids were funerary monuments, i.e. they invariably honored some dead Pharaoh. The Jews invested their engineering progress in building a temple of the Living God instead.
So it stands to reason, in the Hebrews' account of the foreign projects, that immigrants would come in, mess up the project enough, and they would kinda abandon them in progress. But they weren't destroyed.
Well ideally they don't, and when the bubble eventually pops and they go bankrupt we're left with a wide set of useful local models and cheap surplus hardware to run them ourselves. Maybe I'm being overly optimistic.
Yes, I do, but it's not going to prevent me from doing so.
I believe that the current generation of GenAI (as a market, not necessarily as a tech) is going to crash and burn by 2027. I also believe that open-source will stay and will keep helping people and that, as the world becomes more lawless and free-for-all, we'll need all the help we can find.
Me too. Training a neural net of this scale is basically a lossy compression and promting just an interpolation of the data. But explain this to lawyers and courts...
I’ve moved to more closed source projects for this reason (just for the fun of coding rather than sharing). Though I suspect they still use private github repos in their deals to microsoft
Open source is a means to an end, if I can generate an ad-hoc solution that fits my use case more than the often bloated standard open source library for it then I'm all for contributing to the dataset that makes that happen. People who are looking at it from an IP standpoint are less about about solving problems and more about putting themselves on a pedestal of accomplishment just for the sake of getting credit. Especially so for corporations.
It is a bit of a race to the bottom for library components: either you open source and it gets snatched by LLM parties or you keep it closed and good luck selling your wares.
On top of that the open source market will increasingly be flooded with (well intended) AI slop built by junior devs.
Commercial translator services lately are the worst they have ever been. You cant validate that they aren't directly sending your excel with the translation lines into a LLM with no tweaking/checking.
For a indie videogame i work on, we tried a couple translation agencies, and they gave terrible output. At the end, we built our own LLM based agentic translation, with lots of customization for our specific project like building a prompt based on where the menu/string is at, shared glossary, and other features. Testing this against the agencies, it was better because we could customize it for the needs of our specific game.
Even then, at the end of the day, we went with freelancers for some of the languages as we couldn't really validate the AI output on those languages.The freelancers took a month to do the translation vs the 2-3 days we ourselves took for the languages we knew and we could monitor the AI output. But they did a nice job, much better than the agencies.
I feel that what AI really completely kills is those translation services. Its not hard at all to build or customize your own AI system, so if the agency is going to charge you considerable money for AI output, just do it yourself and get a better result. Meanwhile those freelancers are still in demand as they can actually check the project and understand it for a nice translation, unlike the mechanical agencies where you send them the excel and they send it to who knows what or an AI without you being able to check.
I will likely be opensourcing this customizable AI translation system for my project soon.
My mom retired as an independent translator 5 years ago.
She worked freelance 40 years from age 25 to age 65.
In the 15 years that preceded her retirement, she would get less and less work.
Partly she's an introvert who relies on her network to provide work, and her network gradually retired.
But machine translation was the big killer.
Before LLMs, the early versions of Google Translate killed paid translation.
As the market adapted to machine translation, and as the internet became a globalised platform for knowledge work, it also opened up to lower grade translation, and there were suddenly many more translators willing to work at a lower wage.
Prior to Google Translate there were semi-automated systems that would fuzzy match from large databases.
But it'd still rely on a human-in-the-loop for adapting the sentences.
With Google Translate you'd get a super sketchy translation out, very crude and not at all correct or idiomatic in the target language. Any distance between the source and target language (e.g. English -> Chinese) and it'd be one big joke. With plain Google Translate it still is. But the market spoke: Probably you don't need a very good translation most of the time. Especially not if the shitty one is free.
In her later years she moved to transcription of board meetings. She'd type up everything that was said.
I work for a company now that automates transcription via the whisper model and generates summaries that can be adapted by the customer. You pay per minute of transcription, and you can regenerate summaries as much as you want after that until your prompts give the right results.
All of this manual labor that provided for my childhood is gone now.
I couldn't imagine being a professional translator today and not use AI extensively.
But unless I have a legal reason to consult with a professional translator, I probably don't even need one, since LLM-based translation is as good as it gets with just plain LLM usage, and near perfect with automated translation tools that will help you pick both the mood, formality and alternative formulations for your translation.
High-grade translation is massively parallelisable, and a human-in-the-loop is entirely for final proof-reading.
AI translation is so much usable these days I think.
I had a report/article (using Claude) on methods to fall asleep. First method worked amazingly and I am doing that every night. I translated it to Urdu using Gemini (because they are doing translations for a long time, they must be better). First translation used very stiff/dry language, something you might read in a research paper etc. I asked it to simplify it for my mom the language and to my astonishment, it did that very very well.
Urdu surely must not have as much data as other languages. I could not find a single mistake. Yeah there were some weird choices of words in original version, but the simpler was so good it can be published like an article.
It was "cognitive shuffling". While eyes closed, pick a 5-6 letter word (like Blanket), then starting from first letter start making words while visualising each briefly (baloon, bed, bicycle, banana and so on). When first letter is fully exhausted for you, move to next one and do the same.
From what I read and have experienced, I rarely reach second letter. This helped my wife too.
There was another variation where you pick one category (fruits, cities etc) then make words starting from A to Z. Then move to next category. This one requires more mental work for me therefore I haven't tried it yet.
I will take a more positive view, all the translators being put out of buisness will take up literary translation and we will enter a golden age when it comes to translated literature. AI still has a very long ways to go here and there is a great deal of untranslated literature out there and literature which only has a terrible translation. Probably won't happen but maybe the struggling publishing industry will see the opportunity provided by a glut of out of work translators and jump on it.
The premise is absurd. Imagine if coders decided to stop working "on principle" because AI agents "learn from your work"
People would rightly call it out. Somehow gets a pass because its about translation instead of coding?
You've moved to different country with a language that is vastly different from yours. Let's say you're an American moving into the Czech Republic. You need to sign an important document that has legal and business ramifications. Would you trust an AI translation on the document or ask for a professional to be in the loop?
For such a literal case, automatic translations generally suffice. The real translator touch comes about when their is some nuance to the language.
Was that a double entendre or not? If not, you might make a literal translation to get the meaning across. If so, then a literal translation will not get the message across. Vice versa, if it was not a double entendre but you translate it as one, you may confuse the message and if it was and you translate it as such, then the human connection can be maintained.
That is also the tricky bit where you cross from being proficient in the language (say B1-B2) to fluent (C1-C2), you start knowing these double meanings and nuance and can pick up on them. You can also pick up on them when they weren't intended and make a rejoinder (that may flop or land depending on your own skill).
If you are constantly translating with a machine, you won't really learn the language. You have to step away at some point. AI translations present that in full: a translated text with a removed voice; the voice of AI is all of us and that sounds like none of us.
> The real translator touch comes about when their is some nuance to the language.
And as we all know legal language is famous for having no nuance whatsoever, there are no opaque technical terms with hundreds of years of history behind their usage, there is no difference between the legal systems of different countries, and there is no possible difference in case law or the practicalities of legal enforcement. /sarcasm
What is clear to me that in a situation like this neither AI translation nor human translation is sufficient. What the imagined American signing an important legal document in the Czech Republic needs is a lawyer practicing in the Czech Republic who speaks a language the imagined American also speaks.
As someone who has been in that situation before, ‘literal’ translation is not actually a thing. Words and phrases have different meanings between legal systems.
You need a certified translation from someone who is familiar with both legal systems or you’re going to have a very bad time.
Which I think you know from the second part of your statement.
Legal documents likely have much more impact than a random chat with a stranger.
Is the professional _just_ a translator, or an expert in translation _and_ the domain? The latter is preferable; for the former? I'd trust Gemini or Claude.
If the document was some standard thing, like say a standard tax form that every CZ resident needs, then I'd absolutely use Google Translate / AI translation on it. The assumption is that there's no specific trick or catch in the document.
If it was a customized contract then I'd want to use a local legal professional who could also speak English.
When we bought our house in the Netherlands, an official translator was legally required to be present. I'd never have trusted an AI translation for a three-quarters of a million signature.
It's probably advisable to have a lawyer eye through such a document even if that document is in English if there is the slightest question about what it says.
Pacta sunt servanda can be a real bitch sometimes.
You need a real professional for this and not just a translator (the translator can just as well run your document through an LLM and just send you the result). You preferably need a professional who has skin in the game (ie: approved by court).
I don't doubt translators face massive job losses. I hate this "personal interest" style anecdotal news stories though, because it gives me no actual sense/data on how fast/slow this transition is, how widespread etc. It's basically junk food.
for your sins, more anecdote, :)
Guy down the road has used bieng a translator to subsidise his family farm, german to english, and has watched the trade change dramaticly, the bulk(less well paid) now bieng looking over automated translations, some few highly technical things still demand an exactness that cant be trusted to a machine, and some few companys are not concerned with saving every last penny, and prefer , private arangements with humans.Most of the work came through "brokers" but those companies are folding and bieng consodated.The bread and butter is gone, and he is doing other things to make ends meet.It was very good money that could be made from a rural farm house, no so much now, and only because he is very good and has contacts.
Also one of my Aunties did live court translations in Washington, DC, and those will likely still require living humans for the forseable future
The same is happening to a lot of website owners. You lose half or more of your traffic to AI summaries trained on your own content. The cost of producing original content is the same as before.
It makes research harder too, since more and more public information is infected by AI content. Both published posts and internet discussions are tainted.
And then the AI companies threaten to crash the whole economy if we don't let them do it.
> You lose half or more of your traffic to AI summaries trained on your own content.
Wouldn't this be the reason for not calling it transformative but simple copyright theft?
For the millionth time, piracy isn't theft. It's copyright violation, not copyright theft.
No, a reduction in traffic is not sufficient to conclude that a copyright violation has occurred. Sure, it might have. Alternatively it might have produced a lossy summary in which case the reduction in traffic raises some difficult questions about the value of the original work.
In other cases an LLM can synthesize a genuinely useful explanation of a subject that is precisely tailored to the needs of the asker. In those cases the machine output might well prove more useful to the asker than any single original reference would have.
For something like news where what you're paying for is timely delivery it makes sense to restrict automated (not just LLM) access for the first few days because a similarly timely summary will capture the majority of the value proposition of your service.
That's not typical though. For example, I'm certainly not going to be satisfied with a summary of the plot of a book I'm interested in. Would you want to watch a 10 minute highlights reel in place of a 2 hour feature length film?
For the millionth time, the reason we have copyright in the first place is to encourage creation of original creative works. This is clearly stated in the US constitution (and similar phrasing is found in the relevant legal texts of other jurisdictions).
You can apply obsolete legal tests that have been used to enforce this principle all day long, but the central question remains: Does generative AI encourage creation of original creative works?
If the answer is "no", which it clearly is, then whatever laws and legal tests exist to enforce IP rights need to be amended - or the constitution does.
The problem with your definition is that the pre-copyright history is a good example of why we don't need copyright. The US applied copyright laws very late (similar to what China did recently), which led it to be the nation where the citizens read the most in the 19th century. This then led to the cultural explosion we know.
Free reproduction of "original creative works" fuels original creation, too, while creating tight monopolies over intellectual works and universes has led to a decreased creativity around them.
See the dire state of the US film making industry, as an example. Or the vast amount of bizarre lawsuits such as the one for the "Bittersweet Symphony".
I believe that no, copyright does not encourage the creation of creative works, because the mechanism of an exclusive economic monopoly on a creative work supresses more expression than it encourages. Copyright may have hade more validity when the printing press was new, but in a modern context is the wrong mechanism. Instead of encouraging creativity we have instead encouraged capital acquisition and management.
Surely the issue you speak of is largely due to duration? What's wrong with an author (for example) having exclusive rights to a book for 20 or 30 years? Shouldn't that be expected to increase his ability to create additional works?
So if we're talking about the extraordinary power of the state to enforce a right, it has to be in the interest of the public. What is the desirable state we're getting at? Are we maximizing the published creative works, are we concerned with making writing a profession, do we even want culture to be tied to economic activity? Seeing what fans produce with fiction and costumes; I have no doubt that we would have a vibrant and active world of art and fiction without an economic incentive. Do we as a society value the market or the art, and to what degree each?
I agree, but I think there's a second relevant question as well: Does generative AI output original creative works of its own? Obviously the goal here should be maximizing societal benefit. Specifically encouraging the creation of works that we (as a group) find directly useful or otherwise desirable. At least to my mind, human exceptionalism is an explicit non-goal.
I'm already finding the ability of LLMs to synthesize useful descriptions across disparate sources of raw data to be immensely useful. If that puts (for example) scientific textbook authors out of a job I'm not at all sure that would prove to be a detriment to society on the whole. I'm fairly certain that LLMs are already doing better at meeting the needs of the reader than most of the predatory electronic textbook models I was exposed to in university.
> If the answer is "no", which it clearly is,
Why are you so certain of this? It clearly breaks many (most?) of the existing revenue models to at least some extent. But we don't care about the existing revenue models per se. What we care about is long term sustainable creation across society as a whole. So are consumer needs being met in a sustainable manner? Clearly generative AI is (ever increasingly) capable of the former; it's the latter that requires examination.
Summarizing content is not copyright theft.
Normalizing copyright theft is transformative, that's the trick
Normalizing the idea that any such thing as "copyright theft" actually exists would certainly be transformative - and what a stifling mess that would be!
I'm not too fussy about the choice of words. I do the work, they use their monopoly to capture the value. There is no way to opt out, no permission asked. I had a bitter chuckle when Microsoft said that they might "lose social permission" if they can't find good uses for AI.
> It makes research harder too, since more and more public information is infected by AI content.
God, how I hate the billions and billions (sorry Carl Sagan) of pages that look like they're going to have information, but just repeat your question and expand it to 20 or more paragraphs.
"You want to know about how much torque to tighten wheel nuts for a Ford Focus. Torque to tighten wheel nuts is an important aspect in securing the wheels after replacing them...". God damn, it's Eliza being abused to serve the attention economy, and you notice this and wonder if the page would have the correct information or it would just be another bucket of shit pretending to be information... So you think "Let's skip all of that and ask AI...".
Someone, somewhere has to put some information online for the first time. I am one of those people.
This sort of thing was a problem before too. I had a few copycats myself. The difference is that those copycats competed on the same level, and this sort of shallow copying was still too expensive. Google can automate plagiarism at a scale and just shut the competition down.
They are the modern Terminal Railroad Association.
https://en.wikipedia.org/wiki/United_States_v._Terminal_Rail...
Would like to point out that professional translation has been under pressure for much longer than AI.
I have friends that made a descent buck 20-30 years ago translating technical documents like car manuals. Over the years, prices fell from quarters per words to fractions of a cent.
And even though machine translation was barely existent, tools were used to argue higher productivity and therefore lower prices.
This. Google, when it went to a more neural-based translation system about a decade ago, along with DeepL put a serious dent in translation work - particularly around gigs that don't require extremely rigorous accuracy like legal or medical.
Several of my friends who used to work full-time as translators are now supplementing their income with side jobs like foreign language teaching, proofreading, and similar work.
Recently I've noticed that the subtitles on TV shows have become garbage. Often the translations makes no sense in the context of what is happening, but are literal translations of what is being said on screen. I assume it's some kind of AI translation, but who knows. Maybe the translator is just stupid.
To be clear, we had bad subtitles before as well, but that was due to the translators lack of cultural understanding. Most of it made sense at least. Now it's just straight up bad and often makes no sense. Sad!
> I assume it's some kind of AI translation, but who knows. Maybe the translator is just stupid.
AI does a much better job of translating than the stuff I see on TV.
I get the impression that it's done by a lowly paid person who uses a computer dictionary to translate word by word, in a very rushed manner.
> While using AI tools for everyday tasks like finding directions is “low-risk,” human translators will likely need to be involved for the foreseeable future in diplomatic, legal, financial and medical contexts where the risks are “humungous,” according to Benzo
Now it's a classic, you need an expert in order to check the work of the machine, because the "customer" is by definition not able to do it.
Aside from highly technical domain, in purely literary works, I think that the translator is a co-author - maybe IP laws acknowledges that already? I remember the translation of E.A. Poe by C. Baudelaire for instance; I think you could feel Baudelaire's style because it is a lot "warmer" than Poe's. I've also read a translation of a Japanese novel and I was quite disappointed with it. I don't know Japanese but I have read/watched quite a few mangas/animes, so I could sense the speech patterns behind the translations and sometimes thought they could have made better choices.
In any case, one will still need a translator who is good at "prompt engineering" to get a quality translation. I don't know. Maybe translators can add this skill to their CV, so they can propose quick-and-dirty/cheap translations, or no-AI high quality translations.
Some suggest "no-AI" labels on cultural products already - I think if it becomes a reality it will probably act as "quality signaling", because it is becoming more difficult every year to tell the difference between AI and human productions. It won't matter if what you read was written by an AI or a human (if it quacks and looks like a duck...), but what the customer will probably want is to avoid poorly-prompted machine translation.
> it will probably act as "quality signaling", because it is becoming more difficult every year to tell the difference
Note that this only applies to something like a translation where there's some notion of a "correct answer". For other cultural products it's irrelevant (as you say, if it quacks like a duck ...).
Quality signaling is really only necessary in situations where an upfront investment is required and any deception is only revealed sometime later upon use. Safety critical systems such as airbags are a model example of this - a counterfeit of deficient functionality won't be discovered until it deploys, which in most cases will never happen.
That said, while I certainly can't speak to business or diplomatic translations, when it comes to cultural works (ie entertainment) the appeal of machine translation to me has been gradually increasing over time as it gets better. I don't generally find localization desirable and in some cases it even leads to significant confusion when a change somehow munges important details or references. Confusion which I'm generally able to trivially resolve by referencing machine output.
AI translations are getting good. I've been working on our company website in the last month. It's a static website. So, I use Agentic coding tools to implement both content and technical changes. A simple prompt "align the translations with the English version for all pages" and some guard rails in the form of skills and AGENTS.md instructions are doing most of the heavy lifting. To be clear.
These translations are not perfect, yet. But good enough for my needs. Any professional translator services would in any case be beyond our budget. The advantage of using agentic coding tools here (enabled by using a site generator rather than a CMS) is that I can get systematic about dealing with jargon, SEO, and frequently used phrases. I simply document all that and instruct the tool to refer to that. The funny thing is that most of the models are pretty good at fixing their own mistakes if you just ask them too. I asked it to look for examples of "denglish" (German English) in its own German translations and then to fix it. It found a few examples and the suggested fixes were fine.
A lot of people are focusing on the negative here. I like to look at the positive. We're approaching the moment where any person on this planet will be able to communicate directly with any other person on this planet without the need for translators. The tools already exist for this. But they need a lot of work on quality.
A second point here is that the role of English as the most popular intermediary language is disappearing as well. I'm not a native speaker. When I talk to foreigners from wherever, it's mostly in (bad) English. By definition that limits me to talking to people that have had enough education and exposure to English. This is very limiting. A lot of the people we need to talk to here in Germany aren't all that comfortable speaking English.
> A lot of people are focusing on the negative here.
The negative: people are about to lose their jobs.
The positive: AI billionaires become trillionaires.
Why focus on the negative indeed!
And where would you put "billions of people now have access to better translations on demand"?
People talk about business as though only the owners of the business benefit. Everybody else pays the price. But aren't the main beneficiaries all the people using these services?
> billions of people now have access to better translations on deman
As a German speaker, I experience the quality of German language technical documentation steadily declining. 30 years ago, German documentation was usually top notch. With the first machine translations, quality went notably down. Now, with LLM translation, it's often garbage with phrases of obvious nonsense in it.
This is especially true with large companies like IBM, Microsoft or Oracle.
I guess the situation is better for languages where translations only became available with LLM.
The translations aren't better though. Translations across a whole suite of services have got noticeably worse since the advent of AI.
This is explicitly not a benefit to the people using the services.
I think a real life tower of Babel is pretty cool. God destroyed it because he was afraid of a unified humanity.
Douglas Adams's prediction seems more apposite:
"Meanwhile, the poor Babel fish, by effectively removing all barriers to communication between different races and cultures, has caused more and bloodier wars than anything else in the history of creation."
> God destroyed it
Where does it say that? I have never read any Book of Genesis that says "God destroyed the tower".
Also it doesn't say "God was afraid". God doesn't have negative emotions like that. God plans out everything, so He is not "afraid" in the human sense.
In fact, I am fairly certain that the mythical "Tower" for the Jews was sort of a parody of the Pyramids of Egypt and the Ziggurats of Mesopotamia. They were essentially mocking their ancient neighbors in the Levant for such a frivolous project that they believed really didn't honor God, but increased their arrogance and hubris.
In fact, the Sumerians worshipped a god named "Sin" https://en.wikipedia.org/wiki/Sin_(mythology) and it is believed that the "plain of Shinar" and the "wilderness of Sin" are cognate with this term, and therefore represents the ancient deity that was worshipped in that particular case.
For Egypt, the pyramids were funerary monuments, i.e. they invariably honored some dead Pharaoh. The Jews invested their engineering progress in building a temple of the Living God instead.
So it stands to reason, in the Hebrews' account of the foreign projects, that immigrants would come in, mess up the project enough, and they would kinda abandon them in progress. But they weren't destroyed.
But almost everyone speaks English as a second language, and most native English speakers don’t care enough to communicate with everyone.
Por que? Who is ‘everyone’ here?
I mean people outside their immediate surroundings. People rarely seek out connection across cultures. We prefer like minded people.
Well ideally they don't, and when the bubble eventually pops and they go bankrupt we're left with a wide set of useful local models and cheap surplus hardware to run them ourselves. Maybe I'm being overly optimistic.
Anyone else worries that by contributing to open source software these days one is also digging his/her own grave?
Yes. It's one reason I've lost interest in OSS completely.
Yes, I do, but it's not going to prevent me from doing so.
I believe that the current generation of GenAI (as a market, not necessarily as a tech) is going to crash and burn by 2027. I also believe that open-source will stay and will keep helping people and that, as the world becomes more lawless and free-for-all, we'll need all the help we can find.
I’m still convinced that training a model on GPL code makes the model itself a derivative and requires it to be released under GPL terms.
Me too. Training a neural net of this scale is basically a lossy compression and promting just an interpolation of the data. But explain this to lawyers and courts...
I’ve moved to more closed source projects for this reason (just for the fun of coding rather than sharing). Though I suspect they still use private github repos in their deals to microsoft
If you're not sharing the code then what's the benefit of GitHub over self hosted?
Not having to setup and maintain your own self hosted platform. And it also makes it easier to share with specific individuals if you want to.
Free backups
At this point, just use Codeberg and send them a few bucks a month if you want to support them. Fuck GitHub.
Open source is a means to an end, if I can generate an ad-hoc solution that fits my use case more than the often bloated standard open source library for it then I'm all for contributing to the dataset that makes that happen. People who are looking at it from an IP standpoint are less about about solving problems and more about putting themselves on a pedestal of accomplishment just for the sake of getting credit. Especially so for corporations.
I've been collecting my downvotes calling that out ever since the advent of GitHub.
And I don't think most of those came from idealistic people without any vested interest in AI business.
It is a bit of a race to the bottom for library components: either you open source and it gets snatched by LLM parties or you keep it closed and good luck selling your wares.
On top of that the open source market will increasingly be flooded with (well intended) AI slop built by junior devs.
Yes, but we can do Open Source in languages irrelevant to corporates for projects irrelevant to corporates.
The rest is closed source.
Commercial translator services lately are the worst they have ever been. You cant validate that they aren't directly sending your excel with the translation lines into a LLM with no tweaking/checking.
For a indie videogame i work on, we tried a couple translation agencies, and they gave terrible output. At the end, we built our own LLM based agentic translation, with lots of customization for our specific project like building a prompt based on where the menu/string is at, shared glossary, and other features. Testing this against the agencies, it was better because we could customize it for the needs of our specific game.
Even then, at the end of the day, we went with freelancers for some of the languages as we couldn't really validate the AI output on those languages.The freelancers took a month to do the translation vs the 2-3 days we ourselves took for the languages we knew and we could monitor the AI output. But they did a nice job, much better than the agencies.
I feel that what AI really completely kills is those translation services. Its not hard at all to build or customize your own AI system, so if the agency is going to charge you considerable money for AI output, just do it yourself and get a better result. Meanwhile those freelancers are still in demand as they can actually check the project and understand it for a nice translation, unlike the mechanical agencies where you send them the excel and they send it to who knows what or an AI without you being able to check.
I will likely be opensourcing this customizable AI translation system for my project soon.
My mom retired as an independent translator 5 years ago.
She worked freelance 40 years from age 25 to age 65.
In the 15 years that preceded her retirement, she would get less and less work.
Partly she's an introvert who relies on her network to provide work, and her network gradually retired.
But machine translation was the big killer.
Before LLMs, the early versions of Google Translate killed paid translation.
As the market adapted to machine translation, and as the internet became a globalised platform for knowledge work, it also opened up to lower grade translation, and there were suddenly many more translators willing to work at a lower wage.
Prior to Google Translate there were semi-automated systems that would fuzzy match from large databases.
But it'd still rely on a human-in-the-loop for adapting the sentences.
With Google Translate you'd get a super sketchy translation out, very crude and not at all correct or idiomatic in the target language. Any distance between the source and target language (e.g. English -> Chinese) and it'd be one big joke. With plain Google Translate it still is. But the market spoke: Probably you don't need a very good translation most of the time. Especially not if the shitty one is free.
In her later years she moved to transcription of board meetings. She'd type up everything that was said.
I work for a company now that automates transcription via the whisper model and generates summaries that can be adapted by the customer. You pay per minute of transcription, and you can regenerate summaries as much as you want after that until your prompts give the right results.
All of this manual labor that provided for my childhood is gone now.
I couldn't imagine being a professional translator today and not use AI extensively.
But unless I have a legal reason to consult with a professional translator, I probably don't even need one, since LLM-based translation is as good as it gets with just plain LLM usage, and near perfect with automated translation tools that will help you pick both the mood, formality and alternative formulations for your translation.
High-grade translation is massively parallelisable, and a human-in-the-loop is entirely for final proof-reading.
AI translation is so much usable these days I think.
I had a report/article (using Claude) on methods to fall asleep. First method worked amazingly and I am doing that every night. I translated it to Urdu using Gemini (because they are doing translations for a long time, they must be better). First translation used very stiff/dry language, something you might read in a research paper etc. I asked it to simplify it for my mom the language and to my astonishment, it did that very very well.
Urdu surely must not have as much data as other languages. I could not find a single mistake. Yeah there were some weird choices of words in original version, but the simpler was so good it can be published like an article.
Please share the method
It was "cognitive shuffling". While eyes closed, pick a 5-6 letter word (like Blanket), then starting from first letter start making words while visualising each briefly (baloon, bed, bicycle, banana and so on). When first letter is fully exhausted for you, move to next one and do the same.
From what I read and have experienced, I rarely reach second letter. This helped my wife too.
There was another variation where you pick one category (fruits, cities etc) then make words starting from A to Z. Then move to next category. This one requires more mental work for me therefore I haven't tried it yet.
If either works for you, please update :)
I will take a more positive view, all the translators being put out of buisness will take up literary translation and we will enter a golden age when it comes to translated literature. AI still has a very long ways to go here and there is a great deal of untranslated literature out there and literature which only has a terrible translation. Probably won't happen but maybe the struggling publishing industry will see the opportunity provided by a glut of out of work translators and jump on it.
One can dream.
The premise is absurd. Imagine if coders decided to stop working "on principle" because AI agents "learn from your work" People would rightly call it out. Somehow gets a pass because its about translation instead of coding?
You've moved to different country with a language that is vastly different from yours. Let's say you're an American moving into the Czech Republic. You need to sign an important document that has legal and business ramifications. Would you trust an AI translation on the document or ask for a professional to be in the loop?
For such a literal case, automatic translations generally suffice. The real translator touch comes about when their is some nuance to the language.
Was that a double entendre or not? If not, you might make a literal translation to get the meaning across. If so, then a literal translation will not get the message across. Vice versa, if it was not a double entendre but you translate it as one, you may confuse the message and if it was and you translate it as such, then the human connection can be maintained.
That is also the tricky bit where you cross from being proficient in the language (say B1-B2) to fluent (C1-C2), you start knowing these double meanings and nuance and can pick up on them. You can also pick up on them when they weren't intended and make a rejoinder (that may flop or land depending on your own skill).
If you are constantly translating with a machine, you won't really learn the language. You have to step away at some point. AI translations present that in full: a translated text with a removed voice; the voice of AI is all of us and that sounds like none of us.
> The real translator touch comes about when their is some nuance to the language.
And as we all know legal language is famous for having no nuance whatsoever, there are no opaque technical terms with hundreds of years of history behind their usage, there is no difference between the legal systems of different countries, and there is no possible difference in case law or the practicalities of legal enforcement. /sarcasm
What is clear to me that in a situation like this neither AI translation nor human translation is sufficient. What the imagined American signing an important legal document in the Czech Republic needs is a lawyer practicing in the Czech Republic who speaks a language the imagined American also speaks.
The first part is absolutely not true.
As someone who has been in that situation before, ‘literal’ translation is not actually a thing. Words and phrases have different meanings between legal systems.
You need a certified translation from someone who is familiar with both legal systems or you’re going to have a very bad time.
Which I think you know from the second part of your statement.
Legal documents likely have much more impact than a random chat with a stranger.
Is the professional _just_ a translator, or an expert in translation _and_ the domain? The latter is preferable; for the former? I'd trust Gemini or Claude.
If the document was some standard thing, like say a standard tax form that every CZ resident needs, then I'd absolutely use Google Translate / AI translation on it. The assumption is that there's no specific trick or catch in the document.
If it was a customized contract then I'd want to use a local legal professional who could also speak English.
I know many people who until last year would have just signed and hoped for the best. Most people can't afford professional translation services.
When we bought our house in the Netherlands, an official translator was legally required to be present. I'd never have trusted an AI translation for a three-quarters of a million signature.
It's probably advisable to have a lawyer eye through such a document even if that document is in English if there is the slightest question about what it says.
Pacta sunt servanda can be a real bitch sometimes.
These translations usually need to be certified, at least in Germany.
I’d query the top 3 frontier models.
I will totally just do google translate with a phone's cam and that will be it.
You ask the Czech who is handing you the document, cause there's a 99.9% chance they speak perfect English. In Hungary though... yea.
You need a real professional for this and not just a translator (the translator can just as well run your document through an LLM and just send you the result). You preferably need a professional who has skin in the game (ie: approved by court).
I don't doubt translators face massive job losses. I hate this "personal interest" style anecdotal news stories though, because it gives me no actual sense/data on how fast/slow this transition is, how widespread etc. It's basically junk food.
for your sins, more anecdote, :) Guy down the road has used bieng a translator to subsidise his family farm, german to english, and has watched the trade change dramaticly, the bulk(less well paid) now bieng looking over automated translations, some few highly technical things still demand an exactness that cant be trusted to a machine, and some few companys are not concerned with saving every last penny, and prefer , private arangements with humans.Most of the work came through "brokers" but those companies are folding and bieng consodated.The bread and butter is gone, and he is doing other things to make ends meet.It was very good money that could be made from a rural farm house, no so much now, and only because he is very good and has contacts.
Also one of my Aunties did live court translations in Washington, DC, and those will likely still require living humans for the forseable future
[dead]