Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

NY Times copyright suit wants OpenAI to delete all GPT instances

arstechnica.com

Post reply on HN
Print view View original

141–150 of 921 posts

Re: NY Times copyright suit wants OpenAI to delete all GPT instances

#141

Post by thinkingemote »

altals2023 wrote: post #49

Won't hold in court. GPT is a platform mainly providing answer to private individuals asking. Is like you ask a professor a question and he answered verbatim what copyrighted materials available (due to photographic memory) word for word back to you. Now if you take this answer and write a book or publish enmass on blogs for example, then you are the one should be sued by NYT. If GPT use the exact same wordings and p…

The professor having been trained in academia would state the sources of the verbatim quotes. In writing papers he would use references and explicit quotes. There's nothing hidden going on with the professor.

Re: NY Times copyright suit wants OpenAI to delete all GPT instances

#142

Post by visarga »

chmod600 wrote: post #131

Isn't copyright tethered somehow to a notion of "expression"? That is, the same ideas and facts expressed differently are a different work? Sure, when something is clearly derived, or just expressed in a new medium, then I'm sure it's still covered. But if it goes through an LLM and the result bears little resemblance, how can that still fall under copyright?

As you said AI can rewrite articles, obtaining a clean cut separation between ideas and expression. Keep the ideas, write a new text. And if you got multiple sources, the more sources you use the better, it would make the output be even more different. This approach could also check consistency and bias between sources.

Re: NY Times copyright suit wants OpenAI to delete all GPT instances

#144

Post by TeMPOraL »

rich_sasha wrote: post #111

If you forget about the LLM aspect, and simply build a product out of (legally) scraped NYT articles, is that fair use? Let's say I host these, offer some indexing on it, and rewrite articles. Something like, summarise all articles on US-UK relationships over past 5 years. I charge money for it, and all I pay NYT is a monthly subscription fee. To keep things simple, let's say I never regurgitate chunks of verbatim NY…

> Typically I can't take a personal "tier" of a product and charge 3rd parties for derivatives of it. Say like VS Code.

Can't you, though? I'd thought in general, it's a very important for the market to be able to do just that, otherwise everything gets gummed up in webs of exclusive contractual dependencies between established companies.

Re: NY Times copyright suit wants OpenAI to delete all GPT instances

#145

Post by ninjinxo »

hsbauauvhabzb wrote: post #135
jrockway wrote: post #126

I read about this in the Times today (and am surprised that it wasn't on HN already). My guess is that the court will likely find in the Times favor, because the legal system won't be able to understand how training works and because people are "scared" of AI. To me, reading a book, putting it in some storage system, and then recalling it to form future thoughts is fair use. It's what we all do all the time, and I th…

My hard drive can - bit for bit - recall video files. If I serve them to other people on the internet without permission of the copyright holder, that’s called piracy.

But is it still piracy if you compress them and serve only a likeness of the original?

Re: NY Times copyright suit wants OpenAI to delete all GPT instances

#146

Post by visarga »

__loam wrote: post #83
altals2023 wrote: post #49

Won't hold in court. GPT is a platform mainly providing answer to private individuals asking. Is like you ask a professor a question and he answered verbatim what copyrighted materials available (due to photographic memory) word for word back to you. Now if you take this answer and write a book or publish enmass on blogs for example, then you are the one should be sued by NYT. If GPT use the exact same wordings and p…

I hope people start calling out the "well it's fine if a human does it" arguments out for the rat fuck thinking it is. These are computational systems operating at very large scales run by some of the wealthiest companies in the world. If I go fishing, the regulations I have to comply with are very light because the effect I have on the environment is minimal. The regulations for an industrial fishing barge are right…

GPT is like a fleet of small fishing boats, each user driving their boat in another direction, not a fishing barge. For every token written by the model there must be a human who prompted, and then consumed it. It is manual, and personal, and deliberate.

In fact all the demonstrations in the lawsuit PDF were intentionally angling for reproducing copyrighted content. They had to push the model to do it. That won't happen unless users deliberately ask for it. It won't happen en-masse.

Re: NY Times copyright suit wants OpenAI to delete all GPT instances

#148

Post by px43 »

rich_sasha wrote: post #111

If you forget about the LLM aspect, and simply build a product out of (legally) scraped NYT articles, is that fair use? Let's say I host these, offer some indexing on it, and rewrite articles. Something like, summarise all articles on US-UK relationships over past 5 years. I charge money for it, and all I pay NYT is a monthly subscription fee. To keep things simple, let's say I never regurgitate chunks of verbatim NY…

From what I can tell, this has nothing to do with LLMs at all. In the example in the article, the user is asking Bing to go fetch the contents of an article directly from the website, and print it out, which it dutifully does.

Seems like the "problem" is that NYT etc gives privileged access to search engines for indexing their content, but then get upset when snippets of the indexed content is being shown to users without the users having to fight the paywall or whatever.

This article also claims that the screenshot is coming from ChatGPT when it clearly is not.

Re: NY Times copyright suit wants OpenAI to delete all GPT instances

#149

Post by groceryheist »

visarga wrote: post #137
groceryheist wrote: post #122

Earlier quoted context omitted.

The number of unique values in the bloom filter will go up ~exponentially with n. So to control the false positive rate the bloom filter has to grow.

At large enough ngram size there would be very few collisions. You can take for example this text and try in Google with quotes, it won't find anything matching exactly. I tested this 6-gram "it won't find anything matching exactly", no match. Almost anything we write has never been said exactly like that before.

Yes and the fact that the number of unique phrases grows so quickly with n is why the bloom filter needs to grow so that hashed n-grams don't collide.

Re: NY Times copyright suit wants OpenAI to delete all GPT instances

#150

Post by jrockway »

hsbauauvhabzb wrote: post #135
jrockway wrote: post #126

I read about this in the Times today (and am surprised that it wasn't on HN already). My guess is that the court will likely find in the Times favor, because the legal system won't be able to understand how training works and because people are "scared" of AI. To me, reading a book, putting it in some storage system, and then recalling it to form future thoughts is fair use. It's what we all do all the time, and I th…

My hard drive can - bit for bit - recall video files. If I serve them to other people on the internet without permission of the copyright holder, that’s called piracy.

Yeah, but the LLMs can't. They aren't big enough to contain every byte of every NYT article, even with the best-known compression algorithms. Rather, they pick up and remember the same patterns that humans do when they write. Authors of the articles also did that, and so the two algorithms (human writer, LLM inference) end up with the same result. (That doesn't preclude large chunks of text that are actually remembered, though. We humans have large chunks of verbatim text floating around in our brains. Passwords, phone numbers, "I pledge allegiance to the flag...", etc.)

Anyway, like I said, I don't think OpenAI will win this. Someone will produce one verbatim article and the court will make OpenAI pay a bunch of money as though every article could be reproduced verbatim, and AI in the US will be set back that many billion dollars. It probably doesn't matter in the long run; it preserves the status quo for as long as the judge is judging and the newspaper exec is newspaper exec-ing. That's all they need. The next generation will have to figure out how to deal with AI-induced job loss... and climate change. Have fun, next generation!

Post reply on HN