Could somebody explain why this is bad?
I’m not a fan of all this AI stuff.
But I can’t think of an argument besides “Big tech is bad and they should not make money if they use public information to do so.”
I’m genuinely curious. There may be massive amounts of data being processed. But only public data, right? If they can use that data for something, isn’t that something positive? Or at the very least nothing negative?
I always thought anything that is posted in public spaces means making it available for anyone to use anyway. So what am I missing here?
This is more akin to rain water collection up-hill and selling it back to the people downhill
Not really, anyone can go and collect the same water they are collecting. And it’s happening, open source LLMs are quickly catching up and a shit ton of other companies are also crawling the exact same data.
“anyone”. I hate when people use this word knowing full well it’s not true in meaning. “Nothing is stopping you from spending millions of dollars on your own LLM.” Ok.
The web is a bunch of information that is public, sure. People don’t have a reasonable expectation of privacy but they used to have a reasonable expectation that their information would be used in a very specific fashion. Especially in the US where there is a default copyright claim on data. And crawling the web may ignore text that states you can’t use the data. Even if you include a clause saying by accessing the data you agree to the claim. That only works against little people. The “anyone” that can’t actually just go and build a LLM.
Sure but that applies to literally a million other things. There is an absolute ton of shit that companies do that individuals can’t which is still built off of others.
A company can go spend 1B on a new state of the art nuclear reactor which will bring in billions over it’s life time. Will the physicist who discovered the underlying math see any of the profit? No, probably not. And if they do it won’t be nearly a “fair share”. Nor will all the publishers and authors who generated the learning materials that the people working for said company used to build it.
There is tons of public knowledge that can only be utilized with a huge investment, that’s just how a lot of innovation works.
And OpenAI also has a ton of competitors. Sure they have the lead for now but thousands of other companies are also scraping and building LLMs.
You’re not really going to win this argument as I’m an anti-capitalist. So I agree a lot of that stuff is wrong too. I don’t believe you should own other’s labor. The employees should own the company. And I don’t believe in copyright, but it does exist and it’s enforced against individuals, so it’s only fair it’s enforced against them as well. I don’t think you should be allowed to blindly scrape when information could be behind an agreement to use it in a specific manner if accessed. Plus I think it should be opt in based off it being a new use and therefore a new right of copyright. Just as suddenly actors need to worry if they’ll be scanned and owned by a Hollywood studio now. It’s something a reasonable person wouldn’t expect. And that’s why past works are protected from that use.
Things behind a third party privacy policy, sure. You agreed to it, whatever. But your own website? I’m not feeling it.
This comparison is lacking because water is unlike data. The data can still be accessed exactly the same. It doesn’t become less and the access to it is not restricted by other people harvesting it.
While that is true, it does divert peoples’ clicks.
Imagine you wrote a well thought of tech tutorial blog. Is it ok for OpenAI to take your content, train their models, and divert your previous readers away from your blog?
It’s an open ethical question that it’s not straightforward to answer.
EDIT: yes people also learn things and repost them. But the scale at which ChatGPT operates is unprecedented. We should probably let policy catch up. Otherwise we’ll end up with the mess we currently have by letting Google and Facebook collect data for years without restrictions.
it’s not a great comparison I’ll admit, but it’s essentially the same as digital privacy, only one of these is protected in courts and the other is encouraged.
I haven’t sat down to really build a stance on this but it does not sit well.
“public” does not mean you’re allowed to steal it and republish it as a work of your own
That is not what they or LLMs do. And while there is questionable morals around it acting like they are straight up stealing and republishing work hurts having serious discussions about it.
LLMs create statistical distributions of words and phrases based on ingested data, and then sample those distributions given conditional probabilities.
Why should for-profit companies have the right to create these statistical distributions based on our written works without consent? They’re not publishing these distributions, and the purpose of ingesting these texts is not to report on the distributions.
They’re just bottom-trawling the internet and acting as if they have every right to use other peoples’ written works. While people are having “serious discussions” around it, they’re moving forward, ignoring the discussions entirely, and trying to force the conclusion of those discussions to be “well, it’s too late now, anyway”.
Original analysis of public data is not stealing. If it were stealing to do so, it would gut fair use, and hand corporations a monopoly of a public technology. They already have their own datasets, and the money to buy licenses for more. Regular consumers, who could have had access to a corporate-independent tool for creativity and social mobility, would instead be left worse off with fewer rights than where they started.
You should have the discussions first. Not after you already profited off someone else’s work. If the argument should be about whether they can use the data or not, then harvesting it first is absolutely harmful to the discussion you claim is important. You can’t just argue one side is in bad faith when the other side is already objectively acting in bad faith if we are to assume the discussion is real.
But we are way past that. And legally while they are walking a thin line it seems that LLMs are going to win the legal challenges.
I don’t think stopping or slowing LLM development is going to work, because then more questionable countries who really don’t give a fuck about IP will pull ahead.
If you want my honest opinion I don’t think these LLMs companies are stealing and I do think artists are getting the shit end of the stick at the same time. We are heading towards and AI dystopia and I think the way to address is is through more solid social welfare programs instead of fight about IP. While artists are the focus, this AI revolution is coming for all labor. Artists are unfortunately the first ones being impacted by it.
I think people should stop fighting about the minor things and instead prep for the inevitable unemployment this will bring. LLMs are really just the tip of the iceberg.
Yeah, you’re right that it is different from simply stealing content. However the LLMs still use protected material as input and it seems that at least parts of those works can be uniquely identified in the output. That can be considered problematic, even if the data is deconstructed into embeddings inbetween input and output.
As a freelance writer, I write an article for a respected tech website. That article gets views, which in part determines if I get any sort of a performance bonus.
Along comes an AI that scrapes my content, so that when someone asks it a question about how to do “x” on Mac, it spits out an answer based on what it learned from MY article, sometimes regurgitating it word for word, and in doing so deprives me and my publisher of a much need page view.
It affects their revenue, since it affects ad views. It affects my performance bonus.
This isn’t about big tech being “bad”. It’s about writers and other artists not being credited or paid for their work.
This is a good explanation, thank you. I didn’t think about people who literally post stuff to earn money. Since so much talk already revolved around scraping sites like Lemmy, that was all I had in mind.
What you describe sounds like the same problem with services that avoid paywalls or ads of news sites.
In this case I fully aggree that some solution needs to be found.
I don’t consent to my copyrighted material – which is literally everything I write and post online, including this comment – being included in these products. In some cases, I have implicitly consented to allowing this to happen via the EULA of websites I’ve used over the years, but having them actively scraping the web for content means they’re directly bypassing any agreements I may have made with service providers, and that they’re collecting my copyrighted works without my ever having done business of any sort with them.
I haven’t agreed to contribute to their for-profit operation, I’m not being compensated in any way for this participation – whether financially or via the providing of a service – and I don’t believe they have any moral right to decide that I’m going to contribute whether I want to or not.
They’re copying your content, mashing it up with other content, and showing it to their customers, without ever sending their customers to your website. As a result, you don’t get paid and you don’t even get exposure.
Just because something is public, does it mean the source is irrelevant? Not to mention, there’s a lot of stuff that’s not meant to be public that is. A computer won’t know the difference. Public or not, it’s theft to steal the content without credit and monetize it privately.
Let’s said I use AI to write a book, in that case, AI will just grab what someone’s else wrote.
Let’s said I use AI to Write code, AI will just copy someone’s else code.
Let’s said I use AI to make art, AI uses Someone’s else art.
Then, let’s said I sell the book, use the code and make nft’s with the art, since AI “did it” I don’t have to follow any license or give credit to anyone.
About using only public information, that should be an opt in, but instead AI companies are just taking public internet, putting it inside a can a selling it, you like it or not.
Yeah, I don’t really care what they harvest either. I suppose if conversations showed up in chat that would be an issue, but the internet is a public forum anyway and there’s no expectation of privacy here.
If copyright law can work against the individual, it should work against the corporation as well. We can’t only enforce it against the little people. Enforce it for all or for none.
In this instance they’re not even taking copyrighted content. I don’t think random forum posts are copyrightable since they’re not even being reproduced, it’s just being read to create a derivative work.
The expectation that things are not private is totally different from the expectation that things are not being harvested for profit, though. Harvesting things for profit is transforming the public into the private.
Could somebody explain why this is bad? I’m not a fan of all this AI stuff. But I can’t think of an argument besides “Big tech is bad and they should not make money if they use public information to do so.”
I’m genuinely curious. There may be massive amounts of data being processed. But only public data, right? If they can use that data for something, isn’t that something positive? Or at the very least nothing negative? I always thought anything that is posted in public spaces means making it available for anyone to use anyway. So what am I missing here?
If the results were also open and public, it’d be a different conversation.
This is more akin to rain water collection up-hill and selling it back to the people downhill. It’s privatization of a public resource.
Not really, anyone can go and collect the same water they are collecting. And it’s happening, open source LLMs are quickly catching up and a shit ton of other companies are also crawling the exact same data.
“anyone”. I hate when people use this word knowing full well it’s not true in meaning. “Nothing is stopping you from spending millions of dollars on your own LLM.” Ok.
The web is a bunch of information that is public, sure. People don’t have a reasonable expectation of privacy but they used to have a reasonable expectation that their information would be used in a very specific fashion. Especially in the US where there is a default copyright claim on data. And crawling the web may ignore text that states you can’t use the data. Even if you include a clause saying by accessing the data you agree to the claim. That only works against little people. The “anyone” that can’t actually just go and build a LLM.
Sure but that applies to literally a million other things. There is an absolute ton of shit that companies do that individuals can’t which is still built off of others.
A company can go spend 1B on a new state of the art nuclear reactor which will bring in billions over it’s life time. Will the physicist who discovered the underlying math see any of the profit? No, probably not. And if they do it won’t be nearly a “fair share”. Nor will all the publishers and authors who generated the learning materials that the people working for said company used to build it.
There is tons of public knowledge that can only be utilized with a huge investment, that’s just how a lot of innovation works.
And OpenAI also has a ton of competitors. Sure they have the lead for now but thousands of other companies are also scraping and building LLMs.
You’re not really going to win this argument as I’m an anti-capitalist. So I agree a lot of that stuff is wrong too. I don’t believe you should own other’s labor. The employees should own the company. And I don’t believe in copyright, but it does exist and it’s enforced against individuals, so it’s only fair it’s enforced against them as well. I don’t think you should be allowed to blindly scrape when information could be behind an agreement to use it in a specific manner if accessed. Plus I think it should be opt in based off it being a new use and therefore a new right of copyright. Just as suddenly actors need to worry if they’ll be scanned and owned by a Hollywood studio now. It’s something a reasonable person wouldn’t expect. And that’s why past works are protected from that use.
Things behind a third party privacy policy, sure. You agreed to it, whatever. But your own website? I’m not feeling it.
Removed by mod
This comparison is lacking because water is unlike data. The data can still be accessed exactly the same. It doesn’t become less and the access to it is not restricted by other people harvesting it.
While that is true, it does divert peoples’ clicks.
Imagine you wrote a well thought of tech tutorial blog. Is it ok for OpenAI to take your content, train their models, and divert your previous readers away from your blog?
It’s an open ethical question that it’s not straightforward to answer.
EDIT: yes people also learn things and repost them. But the scale at which ChatGPT operates is unprecedented. We should probably let policy catch up. Otherwise we’ll end up with the mess we currently have by letting Google and Facebook collect data for years without restrictions.
it’s not a great comparison I’ll admit, but it’s essentially the same as digital privacy, only one of these is protected in courts and the other is encouraged.
I haven’t sat down to really build a stance on this but it does not sit well.
Explain why copyright exists. If copyright exists, what you state doesn’t hold legal grounds.
Removed by mod
“public” does not mean you’re allowed to steal it and republish it as a work of your own. There are things like copyright and stuff
That is not what they or LLMs do. And while there is questionable morals around it acting like they are straight up stealing and republishing work hurts having serious discussions about it.
LLMs create statistical distributions of words and phrases based on ingested data, and then sample those distributions given conditional probabilities.
Why should for-profit companies have the right to create these statistical distributions based on our written works without consent? They’re not publishing these distributions, and the purpose of ingesting these texts is not to report on the distributions.
They’re just bottom-trawling the internet and acting as if they have every right to use other peoples’ written works. While people are having “serious discussions” around it, they’re moving forward, ignoring the discussions entirely, and trying to force the conclusion of those discussions to be “well, it’s too late now, anyway”.
Original analysis of public data is not stealing. If it were stealing to do so, it would gut fair use, and hand corporations a monopoly of a public technology. They already have their own datasets, and the money to buy licenses for more. Regular consumers, who could have had access to a corporate-independent tool for creativity and social mobility, would instead be left worse off with fewer rights than where they started.
You should have the discussions first. Not after you already profited off someone else’s work. If the argument should be about whether they can use the data or not, then harvesting it first is absolutely harmful to the discussion you claim is important. You can’t just argue one side is in bad faith when the other side is already objectively acting in bad faith if we are to assume the discussion is real.
But we are way past that. And legally while they are walking a thin line it seems that LLMs are going to win the legal challenges.
I don’t think stopping or slowing LLM development is going to work, because then more questionable countries who really don’t give a fuck about IP will pull ahead.
If you want my honest opinion I don’t think these LLMs companies are stealing and I do think artists are getting the shit end of the stick at the same time. We are heading towards and AI dystopia and I think the way to address is is through more solid social welfare programs instead of fight about IP. While artists are the focus, this AI revolution is coming for all labor. Artists are unfortunately the first ones being impacted by it.
I think people should stop fighting about the minor things and instead prep for the inevitable unemployment this will bring. LLMs are really just the tip of the iceberg.
Yeah, you’re right that it is different from simply stealing content. However the LLMs still use protected material as input and it seems that at least parts of those works can be uniquely identified in the output. That can be considered problematic, even if the data is deconstructed into embeddings inbetween input and output.
Thank you. I haven’t thought about copyright just now. This is indeed something that needs to be addressed.
Although I personally still don’t have much of a problem with that. I think copyright laws are highly debatable.
As a freelance writer, I write an article for a respected tech website. That article gets views, which in part determines if I get any sort of a performance bonus.
Along comes an AI that scrapes my content, so that when someone asks it a question about how to do “x” on Mac, it spits out an answer based on what it learned from MY article, sometimes regurgitating it word for word, and in doing so deprives me and my publisher of a much need page view.
It affects their revenue, since it affects ad views. It affects my performance bonus.
This isn’t about big tech being “bad”. It’s about writers and other artists not being credited or paid for their work.
This is a good explanation, thank you. I didn’t think about people who literally post stuff to earn money. Since so much talk already revolved around scraping sites like Lemmy, that was all I had in mind.
What you describe sounds like the same problem with services that avoid paywalls or ads of news sites.
In this case I fully aggree that some solution needs to be found.
Consent.
I don’t consent to my copyrighted material – which is literally everything I write and post online, including this comment – being included in these products. In some cases, I have implicitly consented to allowing this to happen via the EULA of websites I’ve used over the years, but having them actively scraping the web for content means they’re directly bypassing any agreements I may have made with service providers, and that they’re collecting my copyrighted works without my ever having done business of any sort with them.
I haven’t agreed to contribute to their for-profit operation, I’m not being compensated in any way for this participation – whether financially or via the providing of a service – and I don’t believe they have any moral right to decide that I’m going to contribute whether I want to or not.
They can fuck right off.
They’re copying your content, mashing it up with other content, and showing it to their customers, without ever sending their customers to your website. As a result, you don’t get paid and you don’t even get exposure.
Just because something is public, does it mean the source is irrelevant? Not to mention, there’s a lot of stuff that’s not meant to be public that is. A computer won’t know the difference. Public or not, it’s theft to steal the content without credit and monetize it privately.
Let’s said I use AI to write a book, in that case, AI will just grab what someone’s else wrote.
Let’s said I use AI to Write code, AI will just copy someone’s else code.
Let’s said I use AI to make art, AI uses Someone’s else art.
Then, let’s said I sell the book, use the code and make nft’s with the art, since AI “did it” I don’t have to follow any license or give credit to anyone.
About using only public information, that should be an opt in, but instead AI companies are just taking public internet, putting it inside a can a selling it, you like it or not.
Yeah, I don’t really care what they harvest either. I suppose if conversations showed up in chat that would be an issue, but the internet is a public forum anyway and there’s no expectation of privacy here.
If copyright law can work against the individual, it should work against the corporation as well. We can’t only enforce it against the little people. Enforce it for all or for none.
In this instance they’re not even taking copyrighted content. I don’t think random forum posts are copyrightable since they’re not even being reproduced, it’s just being read to create a derivative work.
The expectation that things are not private is totally different from the expectation that things are not being harvested for profit, though. Harvesting things for profit is transforming the public into the private.
Removed by mod