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

651–660 of 921 posts

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

#651

Post by JumpCrisscross »

vidarh wrote: post #273
afavour wrote: post #251

Earlier quoted context omitted.

That would require NYT being willing to sell, which historically they have not been.

I just looked up the share structure; didn't realise the publicly traded shares only appoints 1/3 of the board. Still their second best option is start buying up competitors and going ahead with purging NYT from their training set. That might well end up a worse option for NYT, as they won't stop LLMs from gradually intruding on their space and the moment OpenAI or other LLM providers own major publishers so they don…

> might well end up a worse option for NYT, as they won't stop LLMs from gradually intruding on their space

The Times almost certainly wants its own LLM. I could see them striking a consortium agreement with other newspapers more easily than OpenAI.

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

#652

Post by w4ffl35 »

bonzini wrote: post #483
UrineSqueegee wrote: post #461

Earlier quoted context omitted.

Definition of Transformative Use: The legal concept of transformative use involves significantly altering the original work to create new expressions, meanings, or messages. AI models like GPT don't merely reproduce text; they analyze, interpret, and recombine information to generate unique responses. This process can be argued as creating new meaning or purpose, different from the original works. In the case of the…

Nope, it doesn't work that way. The fact that the LLM can regurgitate original articles doesn't remove the possibility that training can be considered transformative work, or more in general that using copyrighted material for training can be considered fair use. Rather, verbatim reproduction is the proof that copyrighted materials was used. Then the court has to evaluate whether it was fair use. Without verbatim rep…

What if the LLM is running locally and doing all of these things rather than hosted on a webserver which is serving the content?

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

#653

Post by aspenmayer »

tsimionescu wrote: post #326
aspenmayer wrote: post #226

Earlier quoted context omitted.

Would it not logically follow that nothing of value would be lost, even if that were the case? From the point of view of LLMs and content creators, I would treat potential loss of future content being created like I would treat a lost sale. LLMs have value now because of training performed on content that already exists. There must be diminishing returns for certain types of content relative to others. Certain conten…

That's like saying that if a competitor can take your products from your warehouse and sell them for pennies on the dollar, your business has no value. The point is that, to some extent, OpenAI is selling access to NYT content for much cheaper than NYT, while paying exactly 0 to NYT for this content. Obviously, the NYT content costs the NYT more than 0 to produce, so they just can't compete on price with OpenAI, for…

For the sake of argument, let’s say that OpenAI thought it had the rights to process the NYT articles and even display them in part, for the same reasons, fair use or otherwise, that Google can process articles and display snippets of same in its News product, and/or for the same reasons that Google can process books and display excerpts in its Books product. Just like Google in those cases, I would not be surprised to find Google/OpenAI on the receiving end of a lawsuit from rights holders claiming violations of their copyright or IP rights. However, I side with Google then and OpenAI now, as I find both use cases to be fair use, as the LinkedIn case has shown that scraping is fair use. NYT is crying foul because users/consumers of its content archive have derived unforeseen value from said archive and under fair use terms, so NYT has no way to compel OpenAI to negotiate a licensing deal under which they could extract value from OpenAI’s use of NYT data beyond the price paid by any other user of NYT content, whether it be unpaid fair use or fully paid use under license. It feels to me that NYT is engaging in both double-dipping and discriminatory pricing, because they can, and because they’re big mad that OpenAI is more successful than they are with less access to the same or even less NYT data.

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

#654

Post by galangalalgol »

throwuwu wrote: post #367
tkz1312 wrote: post #323

Earlier quoted context omitted.

they might if cliff notes starting copy pasting parts of the source into their articles and passing it off as original writing though :)

The Tolkien estate should get busy suing all the fantasy writers, comic artists, game developers and board and card game companies. Lots of cash there.

They have done some of that actually. Tolkien will be public domain in the nations that are at aithors death+50 in a few days. Sadly, it will be a much longer wait in mine and many others.

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

#655

Post by maxlin »

a_wild_dandan wrote: post #371

The NYT is preparing for a tsunami by building a sandcastle. Big picture, this suit won’t matter, for so many reasons. To enumerate a few: 1. Next gen LLMs will be trained exclusively on “synthetic”/public data. GPT-4V can easily whitewash its entire copyrighted training corpus to be unrecognizably distinct (say reworded by 40%, authors/sources stripped, etc). Ergo there will be no copyright material for GPT-5 to reg…

This is the actual truth. What it sucks for is for citing the data, but GPT-4 doesn't do that to start with unless it's directly from a web result and not the weights.

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

#656

Post by yencabulator »

cyborgx7 wrote: post #635
Robotbeat wrote: post #531

Earlier quoted context omitted.

Correct, just like it’s infringement to reproduce an article from memory using pen and paper intentionally. The person deciding to do that bears responsibility. OpenAI would be liable IFF they were intentionally facilitating that, instead of it being an undesired artifact from overfitting.

I'm pretty sure if you reproduce a work from memory by accident, because you didn't notice your subconscious had just stored the entire article and is now reproducing it word for word, you'd still be guilty of copyright infringement.

The music business is full of examples of that.

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

#657

Post by amadeuspagel »

Two not-so subtle paragraphs about the "partnership" between Microsoft and OpenAI:

> 15. Microsoft Corporation is a Washington corporation with a principal place of business and headquarters in Redmond, Washington. Microsoft has invested at least $13 billion in OpenAI Global LLC in exchange for which Microsoft will receive 75% of that company’s profits until its investment is repaid, after which Microsoft will own a 49% stake in that company.

> 16. Microsoft has described its relationship with the OpenAI Defendants as a “partnership.” This partnership has included contributing and operating the cloud computing services used to copy Times Works and train the OpenAI Defendants’ GenAI models. It has also included, upon information and belief, substantial technical collaboration on the creation of those models. Microsoft possesses copies of, or obtains preferential access to, the OpenAI Defendants’ latest GenAI models that have been trained on and embody unauthorized copies of the Times Works. Microsoft uses these models to provide infringing content and, at times, misinformation to users of its products and online services. During a quarterly earnings call in October 2023, Microsoft noted that “more than 18,000 organizations now use Azure OpenAI Service, including new-to-Azure customers.”

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

#658

Post by pants2 »

bertil wrote: post #623
groceryheist wrote: post #41

The suit demonstrates instances where ChatGTP / Bing Copilot copy from the NYT verbatim. I think it is hard to argue that such copying constitutes "fair use". However, OAI/MS should be able to fix this within the current paradigm: Just learn to recognize and punish plagiarism via RLHF. However, the suit goes far beyond claiming that such copying violates their copyright: "Unauthorized copying of Times Works without p…

Adding an extra constraint of no copying verbatim from a very large and relevant corpus will be hard to guarantee without enormous databases of copyrighted content (which might not be legal to hold) and add an extra objective to a system with many often contradictory goals. I don’t think that’s the technology-sound solution or one in the interest of anyone involved. It’s much more relevant to license content from as…

What if OpenAI were to first summarize or transform the content before training on it? Then the LLM has never actually seen copyrighted content and couldn't produce an exact copy.

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

#659

Post by forrestthewoods »

andy99 wrote: post #645

I don't think the lawsuit has any merit, but I'd still like to encourage Sam Altman et al, if they really care about the greater good, to go Keyser Söze and immediately release torrents of the weights and source code for GPT-4 under GPL.

> I don't think the lawsuit has any merit

The lawsuit fundamentally has merit. It asks a huge open question that no one knows the answer to. The outcome will be extraordinarily impactful. The question must be answered at some point.

The case has merit even if NYT loses across the board.

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

#660

Post by Kon-Peki »

> Because the outputs of Defendants’ GenAI models compete with and closely mimic the inputs used to train them, copying Times works for that purpose is not fair use.

This is interesting. The NYT is specifically saying that the way you use an LLM impacts what you can legally use for training the LLM. They're firing shots at the big guys trying to sell access to an LLM, but not at the little guy self-hosting for fun or academics doing research.

Post reply on HN