Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> then surely a model distilling another model is transformative and fair use.

Yes it is, in the legal/copyright sense of fair use. That's why they ban it in their TOS. Which customers agree to when signing up for the service.



My website’s TOS says not to use it to train AI without permission, yet my website is in the training set of all the big models.

So… my TOS doesn’t matter, but theirs does?


> my TOS doesn’t matter, but theirs does?

Wilhoit Conservatism: In-groups protected by contract law but not bound by it, alongside out-groups bound by contract law, but not protected by it.


Yes - yours is just some optional text nobody reads or understands and is probably not legally required to adhere to. Theirs is a contract signed by their customer who they know did understand it.


Accessing the site may already mean that you agree to TOS. Also if you don’t see an explicit copyright terms on some text on the internet it doesn’t mean that it’s public domain. Same as checking a checkbox. Text being small and somewhere is not an excuse for a corporation to steal and sell other people’s work.


I know that text isn't public domain by default. I'm not talking about pirating IP. I'm talking about reading a website - which is a right that supersedes copyright - even using a computer that learns from it without storing a copy of it or distributing it.

No, you don't need to agree to TOS. I'm sure you never actually agree to them and you're not going to jail for it. If you believe that, I hope you never visited cnn.com, for example. Their Terms of Use is 11,000 words. Are you sure you have agreed to that when you clicked on some random news link? Are you sure you're happy to to surrender your legal defense and its expenses to them if they make a claim against you? You're OK that they have no liability for sharing your PII when they're not authorized to? You won't complain if you pay for a subscription and they don't give you access to the subscription content, don't refund your payment, and don't even tell you why?


That is why every ToS is meaningless unless you can somehow put a gun behind them as the ultimate enforcement with the courts, lawyers and the whole legal system as a fig-leaf-intermediary.


Interesting. So what makes theirs not “optional text nobody reads or understands”?


> what makes theirs not “optional text nobody reads or understands”?

You accept it. You pay consideration for it. If your website has a TOS dickover, that requires someone attest with their legal name and pay you $1, yes, it may be enforceable under some circumstances.


I think you’re failing to understand copyright law.

Without a licence you can’t wget -r a website and make copies of the materials on it, alter them, and so forth.


The term "matters" is proportional to influence. Do you have a team of well financed attorneys?


Your TOS matters insofar as you can prove a person actually read and agreed to it. These are illegal in different ways:

1. Copyright violations (can put you in jail) 2. TOS violations (will be a fine at worst)

Companies do get away with drive-by legal shittiness way too often and frankly the practice needs to be reined in, but at the end of the day the only damages are the financial ones you can prove in court.


And what if the LLM ingested and “understood” it as part of its training?


Well, I guess that's a personal question but the law is pretty clear that only a human being can "understand" anything.


How convenient.

Incidentally, this suggests that once an LLM is capable of accessing and distilling a competitor's LLM without human intervention, then any legal argument about TOS violation is moot. But somehow I doubt that will fly in court.


L1 contracts class: offer, acceptance, and consideration.


The content of the site is subject to licence for making copies. So you’re saying licences don’t matter?

The GPL established this rather clearly. Copyright law doesn’t require consideration.

(The licence itself is a basic BSD licence, so it just requires attribution including in marketing materials, which obviously hasn’t happened.)


> you’re saying licences don’t matter?

Within this context, I don’t think so. I can’t make a website that buries some shrink wrap that requires everyone who reads it become vegan.


So what you are saying is that, if I can somehow get my hands on a copy of Fable, it's fair use to use it to train any models and serve those, since I'm no longer bound by the TOS of the service provider?

Asking for all Anthropic employees who dream big.


How long until books come with TOS, then?


Damn now im looking forward to the day when books end up like physical game disks, where somehow you're not buying the book just a license to it, what a boring dystopia this is lol


As usual, Stallman already predicted this: https://www.gnu.org/philosophy/right-to-read.en.html



Common for college textbooks to have some digital-only component accessed with a one-time key inside the cover. Sometimes time-limited to a single semester.


Have you ever heard of Amazon? Kindle?


Many of them do already and have for decades


Can you please provide a specific example?


Pick up nearly any published book. Turn to the ~3rd page. There will be either a whole page, or sometimes the second half of a page, dedicated to a copyright notice. Very nearly every published book I've ever seen has that identical page. This isn't a recent thing. I grabbed my copy of Diaspora by Greg Egan and opposite the table of contents is a page that starts like this:

Copyright (C) 1998, 2015 by Greg Egan

First Night Shade Books edition 2015

All rights reserved. No part of this book may be reproduced in any manner without the express written consent of the publisher, blah blah (it felt very ironic to transcribe that bit in particular to make this point)


Standard copyright boilerplate. Not terms of service distinct from copyright, which is the subject of this thread.


If you read the whole copyright page of a book and come away with the conclusion that it's anything but terms of use for the book, then we'll just have to agree to disagree.


The thread is about terms of use separate from copyright law. It is the whole premise.


You do not have a contract with the publishers of the book. They are visibly asserting their copyright to deter any defense of ignorance or implied grant of rights to an infringer; but that's not a contract, you did not agree to it before purchasing, there are no contractual terms (scope, duration, faults and compensation, resolution etc.) and nothing in it exceeds the limits the copyright law already sets.

For example, never will you see printed in a book something like "this book is for the exclusive use of the purchaser and you cannot lend, resale or otherwise make available to other parties" - if such a thing was possible, like most software EULAs do, publishers would be all over it.


This is what I meant.


I suspect first-sale doctrine doesn’t allow such.

However, if content is “licensed” instead of “sold” …


They kind of do. There's usually a big scary notice on the imprint page scolding you for even thinking about piracy


We are talking of TOS independent of copyright law.


So far, we have one ruling that says "model distillation by vendor A from vendor B with the intent to use the results to compete with vendor B in vendor B's domain is not fair use". Which makes a degree of sense.

It's possible that distillation for other reasons, with no intent to harm the vendor you distill from, would have been ruled to be fair use. But in law, intent matters.


What was the intent of the original ai companies (anthropic, OpenAI, etc) when they mass-distilled the entire internet to create their training data set?


One could make arguments for OpenAI and Anthropic. But Google Search displays AI results above the SERP - clearly in competition with them. No premise or excuse there.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: