Hacker News Reader

Comment by ronsor | Hacker News Reader

> believes that AI training data is built on the theft of people's labor

I mean, this is an ideological point. It's not based in reason, won't be changed by reason, and is really only a signal to end the engagement with the other party. There's no way to address the point other than agreeing with them, which doesn't make for much of a debate.

> an 1800s plantation owner saying "can you imagine trying to explain to someone 100 years from now we tried to stop slavery because of civil rights"

I understand this is just an analogy, but for others: people who genuinely compare AI training data to slavery will have their opinions discarded immediately.

View on HN

We have clear evidence that millions of copyrighted books have been used as training data because LLMs can reproduce sections from them verbatim (and emails from employees literally admitting to scraping the data). We have evidence of LLMs reproducing code from github that was never ever released with a license that would permit their use. We know this is illegal. What about any of this is ideological and unreasonable? It's a CRYSTAL CLEAR violation of the law and everyone just shrugs it off because technology or some shit.

>We know this is illegal

>It's a CRYSTAL CLEAR violation of the law

in the court of reddit's public opinion, perhaps.

there is, as far as I can tell, no definite ruling about whether training is a copyright violation.

and even if there was, US law is not global law. China, notably, doesn't give a flying fuck. kill American AI companies and you will hand the market over to China. that is why "everyone just shrugs it off".

The "China will win the AI race" if we in the West (America) don't is an excuse created by those who started the race in Silicon Valley. It's like America saying it had to win the nuclear arms race, when physicists like Oppenheimer back in the late 1940s were wanting to prevent it once they understood the consequences.

okay, and?

what do you picture happening if Western AI companies cease to operate tomorrow and fire all their researchers and engineers?

Less slop

China is doing human gene editing and embryo cloning too, we should get right on that. They're harvesting organs from a captive population too, we should do that as well otherwise we might fall behind on transplants & all the money & science involved with that. Lots of countries have drafts and mandatory military service too. This is the zero-morality darwinian view, all is fair in competition. In this view, any stealing that China or anyone does is perfectly fine too because they too need to compete with the US.

All creative types train on other creative's work. People don't create award winning novels or art pieces from scratch. They steal ideas and concepts from other people's work.

The idea that they are coming up with all this stuff from scratch is Public Relations bs. Like Arnold Schwarzenegger never taking steroids, only believable if you know nothing about body building.

Precisely. Nothing is truly original. To talk as though there's an abstract ownership over even an observation of the thing that force people to pay rent to use.. well artists definitely don't pay to whoever invented perspective drawings, programmers don't pay the programming language's creator. People don't pay newton and his descendants for making something that makes use of gravity. Copyright has always been counterproductive in many ways.

To go into details though, under copyright law there's a clause for "fair use" under a "transformative" criteria. This allows things like satire, reaction videos to exist. So long as you don't replicate 1-to-1 in product and purpose IMO it's qualifies as tasteful use.

What the fuck? People also need to pay to access that creative work if the rights owner charges for it, and they are also committing an illegal act if they don't. The LLM makers are doing this illegal act billions of times over for something approximating all creative work in existence. I'm not arguing that creative's make things in a vacuum, this is completely besides the point.

Never heard anything about what you are talking about. There isn't a charge for using tropes, plot points, character designs, etc. from other people's works if they are sufficently changed.

If an LLM reads a free wikipedia article on Aladdin and adds a genie to it's story, what copyright law do you think has been broken?

Meta and Anthropic atleast fed the entire copyrighted books into the training. Not the wikipedia page, not a plot summary or some tropes, they fed the entire original book into training. They used atleast the entirety of LibGen which is a pirated dataset of books.

The central difference is scale.

If a person "trains" on other creatives' works, they can produce output at the rate of one person. This presents a natural ceiling for the potential impact on those creatives' works, both regarding the amount of competing works, and the number of creatives whose works are impacted (since one person can't "train" on the output of all creatives).

That's not the case with AI models. They can be infinitely replicated AND train on the output of all creatives. A comparable situation isn't one human learning from another human, it's millions of humans learning from every human. Only those humans don't even have to get paid, all their payment is funneled upwards.

It's not one artist vs. another artist, it's one artist against an army of infinitely replicable artists.

So this essentially boils down to an efficiency argument, and honestly it doesn't really address the core issue of whether it's 'stealing' or not.

What kind of creative types exist outside of living organisms? People can create award winning novels, but a table do not. Water do not. A paper with some math do not.

What is the basis that an LLM should be included as a "creative type"?

Well a creative type can be defined as an entity that takes other people's work, recombines it and then hides their sources.

LLMs seem to match.

You keep conflating different things.

> We have evidence of LLMs reproducing code from github that was never ever released with a license that would permit their use. We know this is illegal.

What is illegal about it? You are allowed to read and learn from publicly available unlicensed code. If you use that learning to produce a copy of those works, that is enfringement.

Meta clearly enganged in copyright enfringement when they torrented books that they hadn't purchased. That is enfringement already before they started training on the data. That doesn't make the training itself enfringement though.

> Meta clearly enganged in copyright enfringement when they torrented books that they hadn't purchased. That is enfringement already before they started training on the data. That doesn't make the training itself enfringement though.

What kind of bullshit argument is this? Really? Works created using illegally obtained copyrighted material are themselves considered to be infringing as well. It's called derivative infringment. This is both common sense and law. Even if not, you agree that they infringed on copyright of something close to all copyrighted works on the internet and this sounds fine to you? The consequences and fines from that would kill any company if they actually had to face them.

> What kind of bullshit argument is this? Really? Works created using illegally obtained copyrighted material are themselves considered to be infringing as well.

That isn't true.

The copyright to derivative works is owned by the copyright holder of the original work. However using illegaly obtained copies to create a fair use transformative work does not taint your copyright of that work.

> Even if not, you agree that they infringed on copyright of something close to all copyrighted works on the internet and this sounds fine to you?

I agree that they violated copyright when they torrented books and scholarly arguments. I don't think that counts at "close to all copyrighted works on the Internet".

> The consequences and fines from that would kill any company if they actually had to face them.

I don't actually agree that copyright that causes no harm should be met with such steep penalties. I didn't agree when it was being done by the RIAA and even though I don't like facebook, I don't like it here either.

[flagged]

> It's very much based on reason and law.

I have no interest in the rest of this argument, but I think I take a bit of issue on this particular point. I don't think the law is fully settled on this in any jurisdiction, but certainly not in the United States.

"Reason" is a more nebulous term; I don't think that training data is inherently "theft", any more than inspiration would be even before generative AI. There's probably not an animator alive that wasn't at least partially inspired by the works of Disney, but I don't think that implies that somehow all animations are "stolen" from Disney just because of that fact.

Obviously where you draw the line on this is obviously subjective, and I've gone back and forth, but I find it really annoying that everyone is acting like this is so clear cut. Evil corporations like Disney have been trying to use this logic for decades to try and abuse copyright and outlaw being inspired by anything.

It can be based on reason and law without being clear cut - that situation applies to most of reason and law.

> I don't think that training data is inherently "theft", any more than inspiration would be even before generative AI. There's probably not an animator alive that wasn't at least partially inspired by the works of Disney ...

Sure, but you can reason about it, such as by using analogies.

What makes something more or less ideological for you in this context? Is "reason" always opposed to ideology for you? What is the ideology at play here for the critics?

> I mean, this is an ideological point. It's not based in reason

You cant be serious

Reply on news.ycombinator.com