A group of authors filed a lawsuit against Meta, alleging the unlawful use of copyrighted material in developing its Llama 1 and Llama 2 large language models....
They’re the same issue tho. Piracy and using books for corporate AI training both should be fine. The same people going after data freedom are pushing this AI drama too. There’s too much money in copyright holding and it’s not being held by your favorite deviantart artists.
Piracy distributes power. It allows disenfranchised or marginalized people to access information and participate in culture, no matter where they live or how much money they have. It subverts a top-down read-only culture by enabling read-write access for anyone.
Large-scale computing services like these so-called AIs consolidate power. They displace access to the original information and the headwaters of culture. They are for-profit services, tuned to the interests of specific American companies. They suppress read-write channels between author and audience.
One gives power to the people. One gives power to 5 massive corporations.
Also, it’s important to point out that the one that empowers people is the one that is consistently punished far more egregiously.
We have governments blocking the likes of Sci-Hub, Libgen, and Annas-Archive, but nobody is blocking Meta’s LLMs for the same.
If they were treated similarly, I would be far less upset about Meta’s arguments. However it’s clear that governments prioritize the success of business over the success of humanity.
It’s the opposite. Closing down public resources would be regulatory capture and that would be consolidation of power.
Who do you think can afford to pay billions in copyright to produce models? Only mega corporations and pirates. No more small AI companies. No more open source models.
I wish we could be talking about the power imbalances of corporate bodies exercised through the use of capital ownership, instead of squabbling about how that differential is manifested through a specific act of piracy.
The reason we view acts of piracy different when they are committed by corporate bodies is because of the power of their capital, not because the act itself is any different. The issue with Meta and OpenAI using pirated data in the production of LMM’s is that they maintain ownership of the final product to be profited from, not that the LMM comes to exist in the first place (even if it is through questionable means). Had they come to create these models from data that they already owned (I need not remind you that they have already claimed their right to a truly sickening amount of it, without having paid a cent), their profiting from it wouldn’t be any less problematic - LLM’s will still undermine the security of the working class and consolidate wealth into fewer and fewer hands. If we were to apply copyright here as it’s being advocated, nothing fundamental will change in that dynamic; in fact, it will only reinforce the basis of that power imbalance (ownership over capital being the primary vehicle) and delay the inevitable (continued consolidation).
If you’re really concerned with these corporations growing larger and their influence spreading further, then you should be directing your efforts at disrupting that vehicle of influence, not legitimizing it. I understand there’s an enraging double-standard at play here, but the solution isn’t to double down on private ownership, it should be to undermine and seize it for common ownership so that everyone benefits from the advancement.
I wonder if piracy could even benefit these corporations in the long term? Do people who pirate games and movies in their teens and twenties frequently go on to purchase such things when they’re older? I honestly don’t know, but I would love to see a study. I certainly have seen people make that claim.
There you go. Piracy helps. I’m sure game companies and TV producers and so on feel the same way quite often. People who pirate are free marketing for them because they’ll tell other people about the product.
Further, piracy can be reduced or made to not impact you as much if you have the right business model.
Louis CK (before he wrecked his career) famously made millions selling his comedy special through his website for $5 a pop with no Digital Rights Management. You were able to download a copy and keep it forever.
With no DRM, this meant that copies of his special were able to be pirated easily. Prior to releasing this way, he had previously gone on piracy websites and made comments under his pirated specials politely asking people not to pirate, but understanding if they did it because they were too poor.
Despite massive piracy of his special, enough people were happy to pay $5 for a DRM-free copy of his comedy special and if I recall correctly me made $5 million+ on that first special he released like that. It was a massive hit and people were encouraging each other to buy a copy since it was so cheap and respected you as a consumer.
Gabe Newell wasn’t wrong, a big part of piracy always was a service problem.
On December 10, 2011, C.K. released his fourth full-length special, Live at the Beacon Theater. Like Hilarious, it was produced independently and directed by C.K. However, unlike his earlier work, it was distributed digitally on his website, foregoing both physical and broadcast media. C.K. released the special for $5.00 and without DRM, hoping that these factors and the direct relationship between the artist and consumer would effectively deter illegal downloading.
So why are Meta, and say, Sci-Hub are treated so differently? I don’t necessarily disagree, but it’s interesting that we legally attack people who are sharing data altruistically (Sci-Hub gives research away for free so more research can be done, scientific research should be free to the world, because it benefits all of mankind), but when it comes to companies who break the same laws to just make more money, that’s fine somehow.
It’s like trying to improve the world is punished, and being a selfish greedy fucking pig is celebrated and rewarded.
Sci-Hub is so villified, it can be blocked at an ISP level (depending on where you live) and politicians are pushing for DNS-level blocking. Similar can be said for Libgen or Annas-Archive. Is anything like that happening to Meta? No? Huh, interesting. I wonder why Meta gets different treatment for similar behavior.
I am willing to defend Meta’s use of this kind of data after the world has changed how they treat entities like Sci-Hub.Until that changes, all you are advocating for is for corporations to be able to break the law and for altruistic people to be punished. I agree they’re the same, but until the law treats them the same, you’re just giving freebies to giant corporations while fucking yourself in the ass.
To me it always seems to come back to nobility. Big corpo is the new nobility and they have certain privileges not available to the common folk. In theory it shouldn’t exist but in practice it most certainly does.
The aristocracy never died, it just got a new name.
I mean the US is literally built on the fact that the aristocracy in the US didn’t actually want to lose station, so they built a democracy that included many anti-democratic measures from the Senate to the Electoral College to only allowing land-owning white men to vote. The US was purpose built to serve the rich while paying lip-service to the poor.
“Conservatives” were literally always those who wanted to conserve the monarchy and aristocracy. Those were the things they originally wanted to conserve, and plainly still fucking do.
So why are Meta, and say, Sci-Hub treated so differently?
They are not. Meta is being sued, just like Sci-Hub was sued. So, one difference is that the suit involving Meta is still ongoing.
In any case, Meta did not create the dataset. IDK if they even shared it. The researcher who did is also being sued. The dataset has been taken down in response to a copyright complaint. IDK if it is available anywhere anymore. So the dataset was treated just like Sci-Hub. The sharing of the copyrighted material was stopped.
Meta downloading these books for AI training seems fairly straight-forward fair use to me. I don’t see how what Meta did is anything like what Sci-Hub did.
So ISPs are blocking Meta for their breaking of copyright?
Because ISPs block Sci-Hub.
No, one of them is having governments trying to kick off the internet, and the other is allowed to continue doing what they’re doing and the worst they’ll face is a fine. Not even close to the same, completely disproportionate. If they were blocking all Meta LLMs until they had removed all copyrighted material, maybe we could say the same.
ISPs may block sites to prevent unauthorized copying. It’s not a punishment for past wrong-doing. I’m not sure about the details, I think this differs a lot between jurisdictions. But basically, as ISPs they are involved in the unauthorized act of copying. Their servers copy the data to the end user/customer. So, they may be on the hook for infringement themselves if they don’t act.
Again, I am not aware of Meta sharing the copyrighted books in question. So, I don’t know what the legal basis for blocking Meta would be. If ISPs block a site without a legal basis, they are probably on the hook for breach of contract.
IDK on what basis the sharing of Meta’s LLMs could be stopped. If anyone could claim copyright it would be Meta itself and they allow sharing them. (I have doubts if AI models are copyrightable under current US law.)
In its lawsuit Wednesday, the Times accused Microsoft and OpenAI of creating a business model based on “mass copyright infringement,” stating that the companies’ AI systems were “used to create multiple reproductions of The Times’s intellectual property for the purpose of creating the GPT models that exploit and, in many cases, retain large portions of the copyrightable expression contained in those works.”
Publishers are concerned that, with the advent of generative AI chatbots, fewer people will click through to news sites, resulting in shrinking traffic and revenues.
The Times included numerous examples in the suit of instances where GPT-4 produced altered versions of material published by the newspaper.
In one example, the filing shows OpenAI’s software producing almost identical text to a Times article about predatory lending practices in New York City’s taxi industry.
But in OpenAI’s version, GPT-4 excludes a critical piece of context about the sum of money the city made selling taxi medallions and collecting taxes on private sales.
In its suit, the Times said Microsoft and OpenAI’s GPT models “directly compete with Times content.”
If the New York Times’ evidence is true (I haven’t seen the evidence, so I can’t comment on veracity), then you can recreate copyrighted works with LLMs, and as such, they’re doing the same thing as the Pirate Bay, distributing copyrighted works without authorization and making money off the venture.
I expect ISPs would get into a lot of legal trouble if they did.
The NYT sued OpenAI and MS. a) That doesn’t involve Meta. b) It’s a claim by the NYT.
Why should ISPs deny their paying customers access to Meta sites or sites hosting LLMs released by Meta? These customers have contracts with their service providers. On what grounds, would ISPs be in the right to stop providing these internet services?
Both Meta and ChatGPT used books3, it’s functionally the same type of case.
Why should ISPs deny their paying customers access to Meta sites or sites hosting LLMs released by Meta? These customers have contracts with their service providers. On what grounds, would ISPs be in the right to stop providing these internet services?
In the countries where ISP blocking happens, its usually because a copyright holder has sued and demanded blocking at the ISP level and has won in court. Then, the government begins the path of working with ISPs to block the site.
Unless you think most governments that do this do it arbitrarily? No, they do it because a copyright holder sued, like the New York Times has. The NYT has not demanded ISP-level blocking, but that does not mean that they couldn’t. I can’t speak to their choice not to do so other than it seems that companies only save that for truly altruistic groups, and rarely do it for other big corporations.
IDK why you believe this. Breaking contracts is illegal. You get sued and have to pay damages. Some contracts, in some jurisdictions, may allow such arbitrary decisions. In other jurisdictions such clauses may be unenforceable.
altruistic groups
Well, that’s not something that copyright law cares about very much. Unfortunately, this community seems very pro-copyright; very Ayn Rand even. You’re not likely to get much agreement for any sensible reforms; quite the opposite. I don’t think arguing that Meta is doing the same as TPB is going to win anyone over. It’s more likely to get people here to call for more onerous and more harmful IP laws.
Both Meta and ChatGPT used books3, it’s functionally the same type of case.
FWIW, no. the NYT case and this is different in some crucial ways.
“Straight-forward” may be too strong regarding these books. If they inadvertently picked up unauthorized copies while scraping the web, that would definitely not be a problem. That’s what search engines do.
The question is if it is a problem that the researchers knowingly downloaded these copyrighted texts. Owners don’t seem to go after downloaders. IDK if there is case law establishing that the mere act of downloading copyrighted material is infringement. I don’t think there’s anything to suggest that knowing about the copyright status should make a difference in civil law.
In any case, researchers must be able to share copyrighted material, not just for AI training but also any other purpose that needs it. If this is not fair use, then common crawl may not be fair use either. IDK if there is case law regarding the sharing of copyrighted materials as research material, rather than for their content. But I find it hard to see how it could not be fair use, as the alternative would be extremely destructive. So even if the download would normally be infringement, I doubt that it is in this case.
Eventually, we are only talking about a single copy of each book. So, even if researchers were forced to purchase these books, all of AI training would yield only a few extra sales for each title. The benefit to the owners would be very small in relation to the damage to the public.
They’re the same issue tho. Piracy and using books for corporate AI training both should be fine. The same people going after data freedom are pushing this AI drama too. There’s too much money in copyright holding and it’s not being held by your favorite deviantart artists.
It’s not the same issue at all.
Piracy distributes power. It allows disenfranchised or marginalized people to access information and participate in culture, no matter where they live or how much money they have. It subverts a top-down read-only culture by enabling read-write access for anyone.
Large-scale computing services like these so-called AIs consolidate power. They displace access to the original information and the headwaters of culture. They are for-profit services, tuned to the interests of specific American companies. They suppress read-write channels between author and audience.
One gives power to the people. One gives power to 5 massive corporations.
Extremely well-said.
Also, it’s important to point out that the one that empowers people is the one that is consistently punished far more egregiously.
We have governments blocking the likes of Sci-Hub, Libgen, and Annas-Archive, but nobody is blocking Meta’s LLMs for the same.
If they were treated similarly, I would be far less upset about Meta’s arguments. However it’s clear that governments prioritize the success of business over the success of humanity.
It’s the opposite. Closing down public resources would be regulatory capture and that would be consolidation of power.
Who do you think can afford to pay billions in copyright to produce models? Only mega corporations and pirates. No more small AI companies. No more open source models.
I wish we could be talking about the power imbalances of corporate bodies exercised through the use of capital ownership, instead of squabbling about how that differential is manifested through a specific act of piracy.
The reason we view acts of piracy different when they are committed by corporate bodies is because of the power of their capital, not because the act itself is any different. The issue with Meta and OpenAI using pirated data in the production of LMM’s is that they maintain ownership of the final product to be profited from, not that the LMM comes to exist in the first place (even if it is through questionable means). Had they come to create these models from data that they already owned (I need not remind you that they have already claimed their right to a truly sickening amount of it, without having paid a cent), their profiting from it wouldn’t be any less problematic - LLM’s will still undermine the security of the working class and consolidate wealth into fewer and fewer hands. If we were to apply copyright here as it’s being advocated, nothing fundamental will change in that dynamic; in fact, it will only reinforce the basis of that power imbalance (ownership over capital being the primary vehicle) and delay the inevitable (continued consolidation).
If you’re really concerned with these corporations growing larger and their influence spreading further, then you should be directing your efforts at disrupting that vehicle of influence, not legitimizing it. I understand there’s an enraging double-standard at play here, but the solution isn’t to double down on private ownership, it should be to undermine and seize it for common ownership so that everyone benefits from the advancement.
I wonder if piracy could even benefit these corporations in the long term? Do people who pirate games and movies in their teens and twenties frequently go on to purchase such things when they’re older? I honestly don’t know, but I would love to see a study. I certainly have seen people make that claim.
Microsoft famously never went after pirates in Asian countries because despite piracy, it made them the default operating system.
They wanted people to be so used to Windows that they would be willing to pirate it just to use a computer.
It worked and their OS dominance for consumer OSes continues.
There you go. Piracy helps. I’m sure game companies and TV producers and so on feel the same way quite often. People who pirate are free marketing for them because they’ll tell other people about the product.
Further, piracy can be reduced or made to not impact you as much if you have the right business model.
Louis CK (before he wrecked his career) famously made millions selling his comedy special through his website for $5 a pop with no Digital Rights Management. You were able to download a copy and keep it forever.
With no DRM, this meant that copies of his special were able to be pirated easily. Prior to releasing this way, he had previously gone on piracy websites and made comments under his pirated specials politely asking people not to pirate, but understanding if they did it because they were too poor.
Despite massive piracy of his special, enough people were happy to pay $5 for a DRM-free copy of his comedy special and if I recall correctly me made $5 million+ on that first special he released like that. It was a massive hit and people were encouraging each other to buy a copy since it was so cheap and respected you as a consumer.
Gabe Newell wasn’t wrong, a big part of piracy always was a service problem.
So why are Meta, and say, Sci-Hub are treated so differently? I don’t necessarily disagree, but it’s interesting that we legally attack people who are sharing data altruistically (Sci-Hub gives research away for free so more research can be done, scientific research should be free to the world, because it benefits all of mankind), but when it comes to companies who break the same laws to just make more money, that’s fine somehow.
It’s like trying to improve the world is punished, and being a selfish greedy fucking pig is celebrated and rewarded.
Sci-Hub is so villified, it can be blocked at an ISP level (depending on where you live) and politicians are pushing for DNS-level blocking. Similar can be said for Libgen or Annas-Archive. Is anything like that happening to Meta? No? Huh, interesting. I wonder why Meta gets different treatment for similar behavior.
I am willing to defend Meta’s use of this kind of data after the world has changed how they treat entities like Sci-Hub. Until that changes, all you are advocating for is for corporations to be able to break the law and for altruistic people to be punished. I agree they’re the same, but until the law treats them the same, you’re just giving freebies to giant corporations while fucking yourself in the ass.
To me it always seems to come back to nobility. Big corpo is the new nobility and they have certain privileges not available to the common folk. In theory it shouldn’t exist but in practice it most certainly does.
The aristocracy never died, it just got a new name.
I mean the US is literally built on the fact that the aristocracy in the US didn’t actually want to lose station, so they built a democracy that included many anti-democratic measures from the Senate to the Electoral College to only allowing land-owning white men to vote. The US was purpose built to serve the rich while paying lip-service to the poor.
“Conservatives” were literally always those who wanted to conserve the monarchy and aristocracy. Those were the things they originally wanted to conserve, and plainly still fucking do.
How people do not see this is a complete farce.
They are not. Meta is being sued, just like Sci-Hub was sued. So, one difference is that the suit involving Meta is still ongoing.
In any case, Meta did not create the dataset. IDK if they even shared it. The researcher who did is also being sued. The dataset has been taken down in response to a copyright complaint. IDK if it is available anywhere anymore. So the dataset was treated just like Sci-Hub. The sharing of the copyrighted material was stopped.
Meta downloading these books for AI training seems fairly straight-forward fair use to me. I don’t see how what Meta did is anything like what Sci-Hub did.
So ISPs are blocking Meta for their breaking of copyright?
Because ISPs block Sci-Hub.
No, one of them is having governments trying to kick off the internet, and the other is allowed to continue doing what they’re doing and the worst they’ll face is a fine. Not even close to the same, completely disproportionate. If they were blocking all Meta LLMs until they had removed all copyrighted material, maybe we could say the same.
ISPs may block sites to prevent unauthorized copying. It’s not a punishment for past wrong-doing. I’m not sure about the details, I think this differs a lot between jurisdictions. But basically, as ISPs they are involved in the unauthorized act of copying. Their servers copy the data to the end user/customer. So, they may be on the hook for infringement themselves if they don’t act.
Again, I am not aware of Meta sharing the copyrighted books in question. So, I don’t know what the legal basis for blocking Meta would be. If ISPs block a site without a legal basis, they are probably on the hook for breach of contract.
IDK on what basis the sharing of Meta’s LLMs could be stopped. If anyone could claim copyright it would be Meta itself and they allow sharing them. (I have doubts if AI models are copyrightable under current US law.)
https://www.nytimes.com/2024/01/08/technology/openai-new-york-times-lawsuit.html
If the New York Times’ evidence is true (I haven’t seen the evidence, so I can’t comment on veracity), then you can recreate copyrighted works with LLMs, and as such, they’re doing the same thing as the Pirate Bay, distributing copyrighted works without authorization and making money off the venture.
So far, no ISPs are blocking Meta for this.
I expect ISPs would get into a lot of legal trouble if they did.
The NYT sued OpenAI and MS. a) That doesn’t involve Meta. b) It’s a claim by the NYT.
Why should ISPs deny their paying customers access to Meta sites or sites hosting LLMs released by Meta? These customers have contracts with their service providers. On what grounds, would ISPs be in the right to stop providing these internet services?
Both Meta and ChatGPT used books3, it’s functionally the same type of case.
In the countries where ISP blocking happens, its usually because a copyright holder has sued and demanded blocking at the ISP level and has won in court. Then, the government begins the path of working with ISPs to block the site.
Unless you think most governments that do this do it arbitrarily? No, they do it because a copyright holder sued, like the New York Times has. The NYT has not demanded ISP-level blocking, but that does not mean that they couldn’t. I can’t speak to their choice not to do so other than it seems that companies only save that for truly altruistic groups, and rarely do it for other big corporations.
IDK why you believe this. Breaking contracts is illegal. You get sued and have to pay damages. Some contracts, in some jurisdictions, may allow such arbitrary decisions. In other jurisdictions such clauses may be unenforceable.
Well, that’s not something that copyright law cares about very much. Unfortunately, this community seems very pro-copyright; very Ayn Rand even. You’re not likely to get much agreement for any sensible reforms; quite the opposite. I don’t think arguing that Meta is doing the same as TPB is going to win anyone over. It’s more likely to get people here to call for more onerous and more harmful IP laws.
FWIW, no. the NYT case and this is different in some crucial ways.
They pirated the books. Is that not legally relevant?
“Straight-forward” may be too strong regarding these books. If they inadvertently picked up unauthorized copies while scraping the web, that would definitely not be a problem. That’s what search engines do.
The question is if it is a problem that the researchers knowingly downloaded these copyrighted texts. Owners don’t seem to go after downloaders. IDK if there is case law establishing that the mere act of downloading copyrighted material is infringement. I don’t think there’s anything to suggest that knowing about the copyright status should make a difference in civil law.
In any case, researchers must be able to share copyrighted material, not just for AI training but also any other purpose that needs it. If this is not fair use, then common crawl may not be fair use either. IDK if there is case law regarding the sharing of copyrighted materials as research material, rather than for their content. But I find it hard to see how it could not be fair use, as the alternative would be extremely destructive. So even if the download would normally be infringement, I doubt that it is in this case.
Eventually, we are only talking about a single copy of each book. So, even if researchers were forced to purchase these books, all of AI training would yield only a few extra sales for each title. The benefit to the owners would be very small in relation to the damage to the public.