Why Meta’s Project To Translate Automatically Between 200 Languages Will Be Stymied By Copyright
from the copyright-ruins-everything dept
Meta’s AI division has announced two exciting new projects in the field of machine translation:
The first is No Language Left Behind, where we are building a new advanced AI model that can learn from languages with fewer examples to train from, and we will use it to enable expert-quality translations in hundreds of languages, ranging from Asturian to Luganda to Urdu. The second is Universal Speech Translator, where we are designing novel approaches to translating from speech in one language to another in real time so we can support languages without a standard writing system as well as those that are both written and spoken.
The No Language Left Behind technology could have a major impact on how people around the world use the Internet, particularly in the way they access key scientific and medical resources. It would allow people to translate material in one of the more prevalent languages used online, such as English or Spanish, into their own local language once it has been included in the No Language Left Behind project. There’s a crying need for this, for reasons the following Wikipedia article makes clear:
Slightly over half of the homepages of the most visited websites on the World Wide Web are in English, with varying amounts of information available in many other languages. Other top languages are Russian, Spanish, Turkish, Persian, French, German and Japanese.
Of the more than 7,000 existing languages, only a few hundred are recognized as being in use for Web pages on the World Wide Web.
Unfortunately, Meta’s grand vision is unlikely to be realized – because of copyright. Unless online material is released under a permissive license such as the ones devised by Creative Commons, it will be necessary to obtain permission from the copyright holder before a full translation can be made using Facebook’s new tools. It will only take a few high-profile lawsuits from bullying publishers to frighten people away from daring to translate mainstream online articles into their own, poorly-served language without a license.
And so, once again, copyright maximalism will throttle an exciting chance to make the world a better, fairer place by improving access to knowledge – and all to preserve the sanctity of an outdated intellectual monopoly.
Follow me @glynmoody on Twitter, Diaspora, or Mastodon. Originally posted to WalledCulture.
Filed Under: ai, copyright, fair use, translation
Companies: facebook, meta
Comments on “Why Meta’s Project To Translate Automatically Between 200 Languages Will Be Stymied By Copyright”
Much like what already seen in the unprovoked harm the malignancy of copyright maximalism has wantonly inflicted on the blind and deaf communities.
Re:
What [we have] already*
Re:
And the Lakota Tribe.
Re: Re:
It truly was sickening seeing the process of enclosure of commons being applied to a culture’s very language
This comment has been flagged by the community. Click here to show it.
If minimalists didn’t keep trying to destroy copyright, maximalists wouldn’t have been able to get laws that destroyed culture. Blame the minimalists.
Re:
Minimalists provided the excuse, not the reason. Learn your history.
Re:
Copyright maximalism is about the publisher regaining control over what can be published, and destroying any and all means of self publishing is at the top of their list of objectives.
Re: Re:
Copyright maximalism is about the publisher *gaining control over what can be published, actually. It have to have ever belonged to them, thus Disney’s trademarks on the titles of public domain fairy tales and its new policy of changing such names (Tangled instead of Rapunzel, Frozen instead of The Snow Queen) to avoid justified criticism on that basis.
Re: Re:
I’ll try it again, with HTML this time:
Copyright maximalism is about the publisher gaining control over what can be published, actually. It have to have ever belonged to them, thus Disney’s trademarks on the titles of public domain fairy tales and its new policy of changing such names (Tangled instead of Rapunzel, Brave instead of The Bear and the Bow, and Frozen instead of The Snow Queen) to avoid justified criticism on that basis.
Re: Re: Re:
And still some of my comment got cut:
*It doesn’t have to have ever belonged to them…
Re:
You really are the biggest basic bitch for copyright trolls, aren’t you chumley?
Re:
You seem like a moron or gullible, which are you? It’s just a pretext or excuse the copyright cultists use. Not fault of the cultural defenders. This is like blaming the victims. The culture defenders did not give the copyright cultists power. Money and propaganda gave the cultists political power. They leverage money they gain from state-enforced excessive monopoly power into giving them more political power which they then leverage to gain even more state-enforced monopoly power and hence gain more money they then use to bribe the state and corrupt it even more in order to gain more political power. It’s a vicious circle.
You are right in a way. I dont think efforts to minimize or moderate copyrights is best at fighting this circle as I dont think reforms to save the culture will stick because the cultists will leverage their considerable financial resources they have and use their well-developed propaganda and political machine to rollback the reforms. So none of this minimizing or moderating copyright shit for me too.
I’m a copyright abolishing all the way because in my book, copyright is simply too backward, unintelligent, inefficient. It is unjust, and anti-free speech. It gets in way of innovating as in this case. One needs to be able to use other innovations to build on to innovate. One should not have to get permission to innovate. So I’m absolutely fine in seeing copyrights go away as we dont need copyrights hindering innovations. There is no evidence that demonstrate free culture is a threat to innovation or arts so I think its past time to bury this legacy of Queen Anne, byproduct of an unenlightened age, and push for more enlightened, evidence-based innovation promotion policies that is more suited to a digital-based free society, as opposed to policies based on copyright cultist ideology. Policies that is not about serving the greed of an overpriveleged political and economic class, but about the bettering of civilization. You agree?
Re: Re:
If you can’t innovate in a non-infringing way (using the public domain, using ideas, etc.), can you really innovate at all?
Re: Re: Re:
Certainly Copyright make some innovations virtually impossible just like in this case the article talks about and there are many other cases which one can find in old techdirt articles.
I’m thinking right now of another example. Theres many abandoned computer games that could be great but are incomplete or deeply flawed, and its just not possible to get those games completed and improved as the original developers have no interest but other potential innovators have interest in making the games better but can not get permission so no innovation. Many cases,the party with interest in improving or complete the game is not even able to contact the copyright holder as no one knows who to contact or how as copyrights are not always registered. So Copyright fails big time as to innovation in this respect.
I think Copyright is a stupidity which ever breed more stupidities like this or the stupidity in the case that the article talks about.
The copyright cultists are like Copyright is praiseworthy in promoting innovations. Yeah, right, where’s the evidence that Copyright is so useful in promoting innovations in general? It’s plain to me that Copyright do hinder innovations in certain cases. What if Copyright do hinder more than it promotes? I think the premise that Copyright promotes innovations needed to be questioned. If it’s fair to say that Copyright promotes innovations because it promotes some innovations, I think it’s fair to say Copyright does not promote innovations because it hinders some innovations. Can we do better than Copyright as to promote innovations in general? How would we know if we just go by the dogma of the copyright cult rather than go with the evidence? The burden of proof should be on them to prove that Copyright is so great in what it does to society thay it’s worth what its costing the society in progress of innovation. So far, they are failing to prove this so Copyright should go away. Assuming the progress of science and arts is what matter here not democracy values like free speech.
Bluntly, one has to be an idiot to think their ideology is about promoting innovation. It’s about innovations for MONEY for the copyright landlords, not for betterment of society. If the innovations are no good for money for copyright holders, but it betters society, Copyright does not promote. It actively hinders those innovations if it is deemed a threat to the profits of the copyright holders.
If Copyright is indeed about innovations then why do the copyright cult attack the public domain? If you truly care about promoting innovations then how can you expect wealth in innovations from increasingly impoverished public domain? Of course you can try to get permission to use other innovations not in public domain to build your innovation on, but one should not have to have permission to innovate and create value to society. It’s just plainly wrong. I think if one is not depraved with the worship of the Almighty Dollar, is open-minded, truly care about promoting innovations and betterment of society, then one should question the role Copyright plays in our society.
Re: Re: Re:2
Certainly copyright make some innovations virtually impossible just like in this case the article talks about…
You’re shitting me, right? What part of Raziel’s comment pointing to an exception to copyright for machine-created transient copies didn’t you read?
Meta...
“All your translation are belong to us!”
It might be stymied, but there’s no guarantee that it will. Perhaps some bold scientists are willing to risk being thrown in jail or fined but if the research is valuable that the human race should have it, then that research will be free, its revealers not so much.
Take comfort in knowing that there are still brave people in this world
Re:
It’s not even necessarily the research, although i am certain there are plenty of walls there. Thething of it is, automatically translating stuff is a technical violation (the best kind of violation) of copyright.
This is in no small way likely to bring the hammer down on all translation services eventually.
Re: Re:
The thing of it is, automatically translating stuff is a technical violation (the best kind of violation) of copyright.
No it’s not, maximalist.
And so, once again, copyright maximalism will throttle an exciting chance to make the world a better, fairer place by improving access to knowledge – and all to preserve the sanctity of an outdated intellectual monopoly.
Nice fearmongering, but the caselaw says otherwise.
Re:
I’ve done the research and you’re right. The UK also has it that ‘temporary, transient copies’ don’t infringe copyright, but the rule is hard-baked into the legislation rather than relying on interpretation as case law does.
Re: Re:
I’m still dreading the day when they install giant electromagnets at the exits of cinemas to erase the “unauthorised copies” from our memories.
Re: Re: Re:
Don’t worry you’ll be able to license a copy of memories of the film 12-18 month later. Just keep in mind the film may have been re-edited by then, so memories will have no relation to actual events past or present and any persons in those memories are purely fictional, any similarity to real friends is purely coincidental..
Re: Re: Re:
Don’t worry, you can only infringe copyright by reproducing from that memory, which is why it’s subject to an exception.
Raziel is entirely correct.
No-one infringes copyright by visting a website and reading what’s there; similarly someone who knows the foreign language which is in use on a site and translates the contents in his head, does not infringes copyright. Translation software used in the same context is exactly the same.
However pumping the contents of the latest French blockbuster novel through translation software and then publishing the English version without permission would infringe copyright, but the same would apply if a human translator did that, so no change in the law there.
Any one doing the the latter deserves whatever the law throws at them, and that isn’t maximalisation, it’s the current situation.
Taking a more real world example, any website which republished verbatim the whole front page of Le Monde would infringe copyright regardless of whether they translated it into English or left it in the original French.
Re:
Andy J. lays it out nicely and correctly – it’s automation without contravening the law. Considering this under the lights of the Betamax case, this is nothing more than time-shifting becoming “language-shifting”. It’s simply automation being used for convenience, not for profit.
Re:
Any one doing the the latter deserves whatever the law throws at them, and that isn’t maximalisation, it’s the current situation.
I think you mean “maximalism” rather than “maximalization”, but you’re right otherwise.
Re: Re:
maximalism is the philosophy or end state espoused by maximalists; maximalis/zation is the process by which it is achieved. cf capitalist, capitalism and capitalisation. I suspect we live on different sides of the Atlantic hence we may differ in the use of our shared language.
Re: Re: Re:
Maximalization is making as big as possible, it’s maximalism that’s grabbing as much as possible. Similarly, capitalization is the use of capital letters, whereas capitalism is the system of private enterprise. We’re both speaking the same language, you’re just using some words incorrectly.
Re: Re: Re:2
There’s also the type of capitalization that’s involved in capitalizing on an opportunity, but that context is a lot broader than capitalism.
Given how we’ve seen publishers demand payment for the privilege of providing a link sending new customers to view their websites, I cannot envision a scenario where they don’t also demand payment for offering a third party service which then allows people to view those pages in a language which they understand, further increasing traffic to the publishers’ sites.
Maybe I’m missing some nuance here, but doesn’t google translate already do this? Example.
Re:
Beat me to the exact same question.
Dammit, now I look like a copycat — or worse, a brigadier! THIS COMMENT WASN’T HERE WHEN I POSTED, I SWEAR! IT’S THE FILTER’S FAULT, I TELL YOU!
Re:
Part of the problem is that it may not MATTER. Copyright holders have a rather extensive history of suing regardless of the apparent legality of an action, AND of going after the little guys/end-users if the service provider is too large to intimidate by legal threats (like Meta, presumably). Doesn’t do anyone any good if no one wants to use ths service. And I used “apparent” legality because there are very few courts I would be willing to trust would rule in my favor on that issue, even if it is analogous to what Google is already doing. The chilling effect is real, despite the purpose of copyright (in the US at least) being to foster the innovation it usually ends up suppressing.
Re: Re:
Which is why everybody should at least know the basics of the law. It’s a lot easier to fight bullshit if you know it’s bullshit.
Re: Re: Re:
Exactly. Like when I got a letter from my ISP on behalf of RIAA because I ripped the audio from a YouTube vid of the US Marine Band playing Stars and Stripes Forever. I wrote a letter back pointing to where the law said that works of the Federal Government are in the public domain from the moment of creation, and I heard nothing further.
Simple question
Google Translate already allows complete, automatic translation of webpages.
How would Meta be breaking any laws when Google is not?
Several of you ACs are forgetting the second part of the proposition, which is to build and implement an automatic speech translator as well. Google hasn’t said a word about doing this, at least not publicly.
Re:
They automatically generate translateable closed captions on YouTube, which can then be automatically read aloud by a narrator program. That would just combine the existing steps.
Re:
Machine translation of speech is no more an infringement of copyright than machine translation of text. What is at issue is the permanence of the copies, not the medium to be translated.
Last I checked, rules such as for games could be patented but not copyrighted. How exactly can languages – which are nothing but a set of rules for speaking – be copyrighted?
Re:
Well, the last I checked, games absolutely can be copyrighted. How long have you been commenting on this site?
Re: Re:
TBF to Bergman, I think he meant rules for games can’t be copyrighted, not the games themselves, he just worded it really badly.
Re: Re:
Games can be copyrighted, yes.
The specifics, though, are a bit messier.
What you can copyright: Name, logo, original game art (inclusive of concept art, finalized ad pieces, 3d Models, textures if original work, and more), specific implementations of mechanics, game code (again, if the code is original work or an in-house engine, specific tools made just for that game, but NOT the engine if not an in-house engine), any audio work done (original BGM, original SFX, any other original work), levels, original game systems…
This list isn’t exhaustive, btw…
we all know the rules! if something is suggested, let alone brought out, that would be helpful to EVERYONE, those who are such greedy, self-serving assholes will do whatever is necessary to stop it!! aint this a great country, or what??