What Happened to Piracy? Copyright Enforcement Fades as AI Giants Rise
leefang.com
leefang.com
The hypocrisy really grinds my gears.
And a monetary fine is just the cost of doing business if you are big enough.
This is bullshit and you know it. No one would ever want to go to the trouble, expense, and misery of enforcing laws where the so-called victims do not feel wronged and refuse to press complaints to the authorities. And so, copyright enforcement will only ever occur where the rightsholders wish to enforce it. The few lawsuits you see here and there aren't about genuine sentiment of being wrong, but of the rightsholders wanting their cuts. Backroom deals are already being drawn up, and everyone's cool with it.
This is the status quo. If you don't like it, suggest something better, but don't be naive that the law as it stands now could be applied sanely or pragmatically.
SciHub is banned and blocked in several countries, there are default rulings against it in the US etc. Oh, and White House called it "one of the most flagrant notorious market sites in the world".
But when you're an AI company with billions of dollars? Well, your doing it for the good of humanity or something, and of course everything you do is fair use etc.
Plenty of copyright holders don't want their creations to be trained on LLMs, regardless of cut. There is no voice for them.
The general statement of laws being applied differently by size is also more and more obvious in the recent climate.
You think I'm wrong, but if you wrote a song (for instance) and some jackass restaurant was playing it as muzak, ASCAP takes the license money for that. They don't send you a cut. I'd say you have second class rights, but you don't even really have those.
See? Laws are not applied equally.
Somehow their argument that AI models are a derivative work exonerates them from having done piracy in the first place, but not Anna's Archive. It really is about how expensive are your lawyers.
You see this with videogames were they remove Denuvo DRM after a year.
Nah. Media just moved to services, and with that, the main media consumption turned from piracy to subscriptions / ad supported streams.
Lack of public interest is also reflected in the fact that there were not really any development on the piracy front technologically in the last 15 years - or am I behind the times? Between 2000 and 2010, there were a lot more new software, new tech, different ways developed to get files, than there were between 2010 and 2025.
Even with pirating content, I think illegal streaming sites take most of the traffic. There is also much more content than there ever was, making the culture diverse, fragmented, and one-time use only.
So, I think,
1. Services happened to piracy. It's not about collecting files now, but about using an online service to view.
2. At the same time, collecting turned into consuming.
The world's biggest publishing, copyright and IP holders (like Disney, Thomson Reuters, Sony, etc) aren't teenagers or small businesses, and easily have a war chest as big as AI companies, and own about 90% of media IP on which LLMs are trained on, not to mention having lawmakers and artists unions on their side.
If they have a case under current laws, they'll take it to court.
Big AI companies are in a legal blindspot of obvious theft.
- The article you linked
Let me be crystal clear:
Large company pirates stuff: illegal (Anthropic)
Little guy pirates stuff: illegal (though most get away with it)
Large company trains model: legal (Anthropic)
Little guy trains (fine-tunes) model: legal (Civit trainers)
Why the hell are people still repeating the same old shit about "oh so the big guy can ignore laws"? What the fuck is going on?
Also: if copyright enforcement is really getting weaker (no it's not, the article's premise is wrong in the first place), good.
Depending on what your work is, I've probably already pirated it. And I don't stop, it's like a hobby of mine... I've made a point now of downloading every single book title I ever see mentioned (here on HN, elsewhere). Just for shits and giggles. You'll never even know, though, will you? Maybe you're a typographer. I don't pirate fonts, I pirate entire catalogs. Software, games, movies, television, music, literally everything. For years I've been on a search for house blueprints just because (no luck on those, though).
You really are a peasant, and you think that copyright works for you, because theoretically it seems like it should. I guess whatever makes you happy.
Sorry but this is just a competitive reality and the content matters A LOT. Sucks that Elsevier gambled badly on the scientific community putting up with overpriced subscriptions forever, but their concerns can't dictate national policy on this.
And if I'm being honest, I'm tired of the International Brotherhood of Stevedores[0] style of shredding human productivity to protect some special interest group. If Elsevier died tomorrow, we'd lose a curation function to scientific papers, true, but we wouldn't lose the science itself. And while the curation on scientific output is clearly valuable - China is suffering the lack of this while producing prodigious science - I think it's far less important than the scientific output itself. This is especially true of US science.
0: IBS, the AMA, pharmacists, teacher unions, firefighter unions, tax preparers: the distributed cost to society is huge because we decided on protecting these special interest groups. Blocking AI would be a bridge too far.
How reliable is it? Can you just ask an AI for a doi and get a reasonably correct copy of the original article back? Is the level of hallucination induced in science acceptable?
The same applies between companies, by the way, hence the "AI bubble".
The other reason "piracy for AI" is tolerated is because it's not at all clear how to legislate or regulate it. You might think it's a cut and dry case, but lots of other people think the same about the opposite conclusion.
Also, if we're going to bin the entire concept of copyright, can we at least be equal about it? I'd rather not live in a world where humans labor for the remnants of their culture in the content mines while clankers[0] feast on an endless stream of training data.
[0] Fake racial slur for robots or other AI systems.
Nevertheless I thin there is another thing against the LLM training, which is that the scraping seems to be excessive (although it could be made less excessive; there are many ways to help with making it less excessive) and I think it requires too much power (although I don't really know a lot about it).
These are two separate issues, though.
You know, it is really the CC-BY-style most science people care about. Same goes with MIT/BSD open source licenses, while with GPL I suppose it is one the side of CC-BY-SA.