Yeah! I can’t make money running my restaurant if I have to pay for the ingredients, so I should be allowed to steal them. How else can I make money??
Alternatively:
OpenAI is no different from pirate streaming sites in this regard (loosely: streaming sites are way more useful to humanity). If OpenAI gets a pass, so should every site that’s been shut down for piracy.
Which is why they should be legally compelled to publicize all of their datasets, models, research, and share any profits they’ve made with the works they can get provenance data for, because otherwise, it’s an unfair use of the public sphere of content.
One could very easily argue that adblockers are piracy, and those would be stealing from every social media creator, small blog, and independent news site, but I don’t see many people arguing against that, even though that very well includes people who aren’t wealthy corporations.
The issue isn’t necessarily the use of the copyrighted content, it’s the unfair legal stance taken on who can use the content, and how they are allowed to profit (or not profit) from it.
I’m not saying there are no downsides, but I do feel like a simple black and white dichotomy doesn’t properly outline how piracy and generative AI training are relatively similar in terms of who they steal from, and it’s more of a matter of what is done with the content after it is taken that truly matters most.
Generative AI is not going back into the bag. If not OpenAI, then someone else will control it. So we deal with them the next best way, force them to serve us, the people.
Then they can either pay for the copyrighted data they want to train on or lobby for copyright to be reigned in for everyone. Right now, they’re acting like entitled twats with a shit business model demanding they get a free pass while the rest of us would be bankrupted for downloading a Metallica MP3.
The problem isn’t necessarily the use of copyrighted works, (although it can be a problem in many ways) it’s the unfair legal determination of who is allowed to do so.
Google used to act as a directory for the internet along with other web search services. In court, they argued that the content they scrapped wasn’t easily accessible through the searches alone and had statistical proof that the search engine was helping bring people to more websites, not preventing them from going. At the time, they were right. This was the “good” era of Google, a different time period and company entirely.
Since then, Google has parsed even more data, made that data easily available in the google search results pages directly (avoiding link click-throughs), increased the number of services they provide to the degree that they have a conflict of interest on the data they collect and a vested interest in keeping people “on google” and off the other parts of the web, and participated in the same bullshit policies that OpenAI started with their Gemini project. Whatever win they had in the 2000s against book publishers, it could be argued that the rights they were “afforded” back in those days were contingent on them being good-faith participants and not competitors. OpenAI and “summary” models that fail to reference sources with direct links, make hugely inaccurate statements, and generate “infinite content” by mashing together letters in the worlds most complicated markov chain fit in this category.
It turns out, if you’re afforded the rights to something on a technicality, it’s actually pretty dumb to become brazen and assume that you can push these rights to the breaking point.
Google (and search engines in general) is at least providing a service by indexing and making discoverable the websites they crawl. OpenAI is is just hoovering up the data and providing nothing in return. Socializing the cost, privatizing the profits.
That’s not a meaningful distinction, I spent all day using a Copilot search engine because the answers I wanted were scattered across a bunch of different documentation sites.
It was both using the AI models to interpret my commands (not generation at all), and then only publishes content to me specifically.
Depends on what the function was. If the function was to drive ad revenue to your site, then sure, if the function was to get information into the public, then it’s not replacing the function so much as altering and updating it.
If that “altering and updating” means people don’t need to read the original anymore, then it’s not fair use.
TBH I’m for reigning in copyright substantially, and would be on the shitty text generator company side of this, but only if it makes a precedent and erodes copyright as a whole instead of just creating a carveout if you have a lot of moeny for lawyers.
It’s absolutely a meaningful distinction. Search engines push people to tour website where you can capitalize on your audience however you see fit. LLM’s take your content, through them through the mixer and sell it back to people. It’s the difference between a movie reviewer explaining a movie and a dude in an alley selling a pirated copy of the movie.
A) An LLM does not inherently sell you anything. Some companies charge you to run and use their LLMs (OpenAI), and some companies publish their LLMs open source for anyone to use (Meta, Microsoft). With neural chips starting to pop in PCs and phones, pretty soon anyone will be able to run an open source LLM locally on their machine, completely for free.
B) LLMs still rarely regurgitate the exact same original source. This would be more like someone in the back alley putting on their own performance of the movie and morphing it and adjusting it in real time based on your prompts and comments, which is a lot closer to parody and fair use than blatant piracy.
Yeah! I can’t make money running my restaurant if I have to pay for the ingredients, so I should be allowed to steal them. How else can I make money??
Alternatively:
OpenAI is no different from pirate streaming sites in this regard (loosely: streaming sites are way more useful to humanity). If OpenAI gets a pass, so should every site that’s been shut down for piracy.
If OpenAI wants a pass, then just like how piracy services make content freely open and available, they should make their models open.
Give me the weights, publish your datasets, slap on a permissive license.
If you’re not willing to contribute back to society with what you used from it, then you shouldn’t exist within society until you do so.
Piracy steals from the rich and gives to the poor. ChatGPT steals from the rich and the poor and keeps for itself.
Which is why they should be legally compelled to publicize all of their datasets, models, research, and share any profits they’ve made with the works they can get provenance data for, because otherwise, it’s an unfair use of the public sphere of content.
One could very easily argue that adblockers are piracy, and those would be stealing from every social media creator, small blog, and independent news site, but I don’t see many people arguing against that, even though that very well includes people who aren’t wealthy corporations.
The issue isn’t necessarily the use of the copyrighted content, it’s the unfair legal stance taken on who can use the content, and how they are allowed to profit (or not profit) from it.
I’m not saying there are no downsides, but I do feel like a simple black and white dichotomy doesn’t properly outline how piracy and generative AI training are relatively similar in terms of who they steal from, and it’s more of a matter of what is done with the content after it is taken that truly matters most.
No they shouldn’t. They should cease to exist
Generative AI is not going back into the bag. If not OpenAI, then someone else will control it. So we deal with them the next best way, force them to serve us, the people.
It probably will, though, once model collapse sets in.
That’s the irony, really… the more successful it is, the sooner it’ll poison itself to death.
Then they can either pay for the copyrighted data they want to train on or lobby for copyright to be reigned in for everyone. Right now, they’re acting like entitled twats with a shit business model demanding they get a free pass while the rest of us would be bankrupted for downloading a Metallica MP3.
I think this better solves the issue.
The problem isn’t necessarily the use of copyrighted works, (although it can be a problem in many ways) it’s the unfair legal determination of who is allowed to do so.
Nobody should profit from copyright violation. Yes, copyright law needs to change, but making money isn’t an exception
Good luck putting the cat back in the bag.
Well if everyone who’s copyrighted work independently sues OpenAI, that cat will be deceased real quick due to bankruptcy
Fuck copyright they used gplv3 code why isnt it open source
I have cats. Putting them back in a bag or box is easier
K, so Google should be shut down too?
They can’t operate without scraping copyrighted data.
This is a false equivalency.
Google used to act as a directory for the internet along with other web search services. In court, they argued that the content they scrapped wasn’t easily accessible through the searches alone and had statistical proof that the search engine was helping bring people to more websites, not preventing them from going. At the time, they were right. This was the “good” era of Google, a different time period and company entirely.
Since then, Google has parsed even more data, made that data easily available in the google search results pages directly (avoiding link click-throughs), increased the number of services they provide to the degree that they have a conflict of interest on the data they collect and a vested interest in keeping people “on google” and off the other parts of the web, and participated in the same bullshit policies that OpenAI started with their Gemini project. Whatever win they had in the 2000s against book publishers, it could be argued that the rights they were “afforded” back in those days were contingent on them being good-faith participants and not competitors. OpenAI and “summary” models that fail to reference sources with direct links, make hugely inaccurate statements, and generate “infinite content” by mashing together letters in the worlds most complicated markov chain fit in this category.
It turns out, if you’re afforded the rights to something on a technicality, it’s actually pretty dumb to become brazen and assume that you can push these rights to the breaking point.
Google (and search engines in general) is at least providing a service by indexing and making discoverable the websites they crawl. OpenAI is is just hoovering up the data and providing nothing in return. Socializing the cost, privatizing the profits.
Uh, that’s objectively false.
OoenAI also provides ChatGPT as a “free” service, and Google has made billions off of that “free” service they oh so altruistically provide you.
Google points to your content so others can find it.
OpenAI scrapes your content to use to make more content.
That’s not a meaningful distinction, I spent all day using a Copilot search engine because the answers I wanted were scattered across a bunch of different documentation sites.
It was both using the AI models to interpret my commands (not generation at all), and then only publishes content to me specifically.
Technically it is meaningful, fair use is for specifically things that don’t replace the original in function.
Depends on what the function was. If the function was to drive ad revenue to your site, then sure, if the function was to get information into the public, then it’s not replacing the function so much as altering and updating it.
If that “altering and updating” means people don’t need to read the original anymore, then it’s not fair use.
TBH I’m for reigning in copyright substantially, and would be on the shitty text generator company side of this, but only if it makes a precedent and erodes copyright as a whole instead of just creating a carveout if you have a lot of moeny for lawyers.
I’m talking about the training phase of LLMs.that is the portion that is doing the scraping and generation of copy written data.
You using an already trained LLM to do some searches is not the same thing.
It’s absolutely a meaningful distinction. Search engines push people to tour website where you can capitalize on your audience however you see fit. LLM’s take your content, through them through the mixer and sell it back to people. It’s the difference between a movie reviewer explaining a movie and a dude in an alley selling a pirated copy of the movie.
A) An LLM does not inherently sell you anything. Some companies charge you to run and use their LLMs (OpenAI), and some companies publish their LLMs open source for anyone to use (Meta, Microsoft). With neural chips starting to pop in PCs and phones, pretty soon anyone will be able to run an open source LLM locally on their machine, completely for free.
B) LLMs still rarely regurgitate the exact same original source. This would be more like someone in the back alley putting on their own performance of the movie and morphing it and adjusting it in real time based on your prompts and comments, which is a lot closer to parody and fair use than blatant piracy.
This is actually a very good comparison because restaurants use this argument all the time, except for wages:
“I can’t make money running my restaurant if I have to pay a living wage to my servers, so you should pay them with tips. How else can we stay open?”
These business that can’t operate profitably like any other business should fail.