The New York Times is suing OpenAI and Microsoft for copyright infringement • Hacker News Board

Re: The New York Times is suing OpenAI and Microsoft for copyright infringement

#841

Post by malwrar » Thu, Dec 28, 2023, 7:25 AM UTC

kbos87 wrote: post #708

whywhywhywhy wrote: post #644

Earlier quoted context omitted.

The world you’re hoping for will put all AI tech only within the hands of the established top 10 media entities, who traditionally have never compensated fairly anyway. Sorry but if that’s the alternative to some writers feeling slighted, I’ll choose for the writers to be sad and the tech to be free.

“Feeling slighted” is a gross understatement of how a lack of compensation flowing to creators has shaped the internet and the wider world over the past 25 years. If we have a problem with the way top media companies compensate their creators, that is a separate issue - not a justification for layering another issue on top.

I think the introduction of an expectation for compensation has generally brought down the quality of content online. Different people and incentives appear to get involved once content == money, vs content == creative expression.

Re: The New York Times is suing OpenAI and Microsoft for copyright infringement

#842

Post by rajamaka » Thu, Dec 28, 2023, 7:36 AM UTC

015a wrote: post #831

whichfawkes wrote: post #572

Earlier quoted context omitted.

Human brains are still the main legal agents in play. LLMs are just a computer programs used by humans. Suppose I research for a book that I'm writing - it doesn't matter whether I type it on a Mac, PC, or typewriter. It doesn't matter if I use the internet or the library. It doesn't matter if I use an AI powered voice-to-text keyboard or an AI assistant. If I release a book that has a chapter which was blatantly cop…

> Human brains are still the main legal agents in play. No, they're not. This is The New York Times (a corporation) vs OpenAI and Microsoft (two more corporations).

Aren't corporations considered 'persons' in the US?

Re: The New York Times is suing OpenAI and Microsoft for copyright infringement

#843

Post by happymellon » Thu, Dec 28, 2023, 7:48 AM UTC

logicchains wrote: post #650

happymellon wrote: post #647

Earlier quoted context omitted.

> while humans face no such restriction. I have no idea what on earth you are talking about. People and corporations are sued for copyright infringement all the time. https://copyrightalliance.org/copyright-cases-2022/ Reading and consuming other people content isn't illegal, but it also wouldn't be for a computer. Reading and consuming content with the sole purpose of reproducing it verbatim is frowned upon, and can…

>I have no idea what on earth you are talking about. People and corporations are sued for copyright infringement all the time. They're sued for _producing content_, not consuming content. If a human takes copyrighted output from an LLM and publishes it, they're absolutely liable if they violated copyright. >Reading and consuming other people content isn't illegal, but it also wouldn't be for a computer. That is absol…

> That's irrelevant here because people training LLMs aren't feeding them copyrighted content for the sole purpose of reproducing it verbatim.

Disagree, it is completely relevant when discussing computers Vs people, the bar that has already been set is alternative uses.

LLMs don't have a purpose outside of regurgitating what it has ingested. CD burners at least could be claimed they were backing up your data.

Re: The New York Times is suing OpenAI and Microsoft for copyright infringement

#844

Post by CaptainFever » Thu, Dec 28, 2023, 8:22 AM UTC

blagie wrote: post #748

CaptainFever wrote: post #547

Earlier quoted context omitted.

Copying is not theft. Stealing a thing leaves one less left Copying it makes one thing more; that’s what copying’s for.

My code was AGPL. OpenAI can go to h..l (Footnote: I like your poem. It conveys the concept much better than anywhere I'd ever seen before)

Thanks, but it's not my poem! You can find it here: https://blog.ninapaley.com/2009/12/15/minute-meme-1-copying-...

Re: The New York Times is suing OpenAI and Microsoft for copyright infringement

#845

Post by thiht » Thu, Dec 28, 2023, 9:31 AM UTC

logicchains wrote: post #624

DamnInteresting wrote: post #601

I have deeply mixed feelings about the way LLMs slurp up copyrighted content and regurgitate it as something "new." As a software developer who has dabbled in machine learning, it is exciting to see the field progress. But I am also an author with a large catalog of writings, and my work has been captured by at least one LLM (according to a tool that can allegedly detect these things). Overall, current LLMs remind me…

>All that aside, I tend to agree with the hypothesis that LLMs are a fad that will mostly pass. For professionals, it is really hard to get past hallucinations and the lack of citations. For writers maybe, but absolutely not for programmers, it's incredibly useful. I don't think anyone who's used GPT4 to improve their coding productivity would consider it a fad.

Copilot has been way more useful to me than GPT4. When I describe a complex problem where I want multiple solutions to compare, GPT4 is useless to me. The responses are almost always completely wrong or ignore half of the details I’ve written in the prompt. Or I have to write them with already a response in mind, which kinda defeats why I would use it in the first place.

Copilot provides useful autocompletes maybe… 30% of the time? But it doesn’t waste too much as it’s more of a passive tool.

Re: The New York Times is suing OpenAI and Microsoft for copyright infringement

#846

Post by edwintorok » Thu, Dec 28, 2023, 9:50 AM UTC

TheCoreh wrote: post #640

tsimionescu wrote: post #610

Earlier quoted context omitted.

Conceptually, it wouldn't be very hard to take the candidate output and run it through a text matching phase to see if there are ~exact matches in the training corpus, and generate other output if there are (probably limited to the parts of the training corpus where rights couldn't be obtained normally). Of course, it would be quite compute heavy, so it would add significantly to the cost per query.

GitHub Copilot supports that: https://docs.github.com/en/copilot/configuring-github-copilo... Given how cheap text search is compared with LLM inference, and that GitHub reuses the same infrastructure for its code search, I doubt it adds more than 1% to the total cost.

It is questionable whether that filtering mechanism works, previous discussion: https://news.ycombinator.com/item?id=33226515

But even if it did an exact match search is not enough here. What if you take the source code and rename all variables and functions? The filter wouldn't trigger, but it'd still be copyright infringement (whether a human or a machine does that).

For such a filter to be effective it'd at least have to build a canonical representation of the program's AST and then check for similarities with existing programs. Doing that at scale would be challenging.

Wouldn't it be better to: * Either not include copyrighted content in the training material in the first place * Explicitly tag the training material with license and origin infornation, such that the final output can produce a proof of what training material was relevant for producing that output and don't mix differently licensed content.

Re: The New York Times is suing OpenAI and Microsoft for copyright infringement

#847

Post by edwintorok » Thu, Dec 28, 2023, 9:53 AM UTC

They are not the only ones to sue. There is also a class action: https://githubcopilotlitigation.com/

Re: The New York Times is suing OpenAI and Microsoft for copyright infringement

#848

Post by edwintorok » Thu, Dec 28, 2023, 10:04 AM UTC

andy99 wrote: post #767

twoodfin wrote: post #754

Earlier quoted context omitted.

There are two problems with the “kid” analogy: a) In many closely comparable scenarios, yes, it’s copyright infringement. When Francis Ford Coppola made The Godfather film, he couldn’t just be “inspired” by Puzo’s book. If the story or characters or dialog are similar enough, he has to pay Puzo, even if the work he created was quite different and not a literal “copy”. b) Training an LLM isn’t like giving someone a bo…

> This copy is not a transitory copy in service of a fair use Training is almost certainly fair use, so it's exactly a transitory copy in service of fair use. Training, other than the brief "transitory copy" you mention is not copying, it's making a minuscule algorithmic adjustment based on fleeting exposure to the data.

If you overtrain the model may include verbatim copies of your training material, and may be able to produce verbatim copies of the original in its output.

If Microsoft truly believes that the trained output doesn't violate copyright then it should be forced to prove that by training it on all its internal source code, including Windows.

Re: The New York Times is suing OpenAI and Microsoft for copyright infringement

#849

Post by noobermin » Thu, Dec 28, 2023, 10:04 AM UTC

iandanforth wrote: post #180

Aurornis wrote: post #20

The arguments about being able to mimic New York Times “style” are weak, but the fact that they got it to emit verbatim NY Times content seems bad for OpenAI: > As outlined in the lawsuit, the Times alleges OpenAI and Microsoft’s large language models (LLMs), which power ChatGPT and Copilot, “can generate output that recites Times content verbatim

Arguing whether it can is not a useful discussion. You can absolutely train a net to memorize and recite text. As these models get more powerful they will memorize more text. The critical thing is how hard is it to make them recite copyrighted works. Critically the question is, did the developers put reasonable guardrails in place to prevent it? If a person with a very good memory reads an article, they only violate…

I hate to do this but this then becomes a "only bad people with a gun kill people" argument. Even most but the most ardent gun rights advocates in that scenario think they shouldn't be extended to very powerful weapons like bombs or nuclear weapons. In this situation then, this logic would be "sure this item allows a person to kill thousands or millions of people, but really the only person at fault in such a situation is the one who presses the button." This ignores the harm done and only focuses on who gets the fault, as if all discourse on law is determining who is a bad guy or a good guy in a movie script.

The general prescription (that I do agree not everyone accepts) society has come up with is we relegate control of some of these weapons to governments and outright ban others (like chemical weapons, biological weapons, and such) through treaties. If LLMs can cause so much damage and their use can be abused so widely, you have to stop focusing on questions about whether a user is culpable or not and move to consider whether their wide use is okay and shouldn't be controlled.

Re: The New York Times is suing OpenAI and Microsoft for copyright infringement

#850

Post by tsimionescu » Thu, Dec 28, 2023, 11:05 AM UTC

denton-scratch wrote: post #753

tsimionescu wrote: post #683

Earlier quoted context omitted.

I very much doubt the company or foundation you were working for was selling the non-copyrighted software. If it was, it probably only worked on very specific hardware that you also produced and were selling, and thus copying it was largely useless. If you were working for a university, than the university obviously doesn't make money from selling software, and thus doesn't care for copyright as much. Also, craftsmen…

> I very much doubt the company [...] was selling the non-copyrighted software Well you'd be mistaken. Lately, it was custom software, for a particular client, and of no interest to others. Earlier, it was before software copyright was a thing, and computer manufacturers gave software away to sell the hardware. At the very beginning, yes, it was "very specific" hardware; it was Burroughs hardware, which used Burrough…

> Well you'd be mistaken. Lately, it was custom software, for a particular client, and of no interest to others. Earlier, it was before software copyright was a thing, and computer manufacturers gave software away to sell the hardware.

Then I am not mistaken: the company was initially selling hardware, with the software being just a value add as you say (no copyright: no interest in trying to sell, exactly my point). Then, you were being paid for building software that (a) was probably not being made public anyway, and (b) would not have been of interest to others even if it were.

Even so, if someone came to your client and offered to take on the software maintenance for a much lower price, you might have lost your client entirely. This has very much happened to contractors in the past.

And my point is you couldn't have a Microsoft or Adobe or possibly even RedHat if you didn't have copyright protecting their business. So, you'd probably not have virtually any kind of consumer software.