Meta has scraped data from the most-trafficked domains on the internet —including news organizations, education platforms, niche forums, personal blogs, and even revenge porn sites—to train its artificial intelligence models, according to a leaked list obtained by Drop Site News.
By scraping data from roughly 6 million unique websites, including 100,000 of the top-ranked domains, Meta has generated millions of pages of content to use for Meta’s AI-training pipeline.
The sites that Meta scrapes consist of copyrighted content, pirated content, and adult videos, some of whose content is potentially illegally obtained or recorded, as well as news and original content from prominent outlets and content publishers.
They include mainstream businesses like Getty Images, Shopify, Shutterstock, but also extreme pornographic content, including websites advertising explicit sexual content and humiliation porn that exploits teenagers.
Sounds pretty god damned illegal.
It’s illegal for you and me. Didn’t you get that by now?
Outright piracy? It’s not allowed, but it’s supposed to be a civil matter.
Videos posted without permission? I don’t think the audience is liable for that.
Scraping despite robots.txt? If that’s illegal for its own sake, then it’s overreaching on ‘unauthorized access.’
Training on any of this? … nah, it’s probably fine.
A pile of linear algebra that knows what pornography looks like does not serve the same function as any particular example. No more than one video infringes on another for the general idea of cameras pointed at naked people. Producing the same kind of thing is not infringement. (Though if it involves Shrek, the trademark people will have angry and confusing questions.)
Reproducing any particular input is a failure of training. Even the Bible should be paraphrased past about Genesis 1:9. The whole idea is getting the vibe of everything we’ve ever published. Cliff notes, passable imitation of the writing style, couple passages everyone’s quoted verbatim.
An encyclopedia article about a book doesn’t become illegal if we learn the author shoplifted it.
And nothing is going to happen to Meta because when they take your property, the laws don’t apply to them.
This is a feature of the system, it just now becoming obvious for the average peon.
If I scrape Meta’s AI to develop my own, would that be fair game? I’m genuinely curious about the legality of this.
Tehnically you would be breaking terms of service and license, but in a legal sense we don’t know if that would be enforceable. Still hasn’t been answered by courts.
So far, OpenAI, anthropic et al hasn’t sued anyone over it, but they have cut account access when it’s discovered to be used for that purpose
It’s how early versions of deepseek were trained iirc, it’s called distillation
Lemmy really hates piracy… in this specific context.
And a lot of the extreme and extremist content going into these things is just Twitter. People post all kinds of shit from all kinds of places. At what point is this like clutching pearls over what the Internet Archive has saved? They’re trying to grab anything you could see.
It’s not some hacking and exfiltration campaign. Meta’s just bad at spidering. How do you go breadth-first across the entire internet and still DDoS any particular site? You don’t decide to check every DeviantArt account, at the same time, you dolts.
Lemmy really hates piracy… in this specific context.
Specifically, Lemmy hates it when corporations profit by using people’s work without permission or payment, especially at a large scale.
I don’t think Lemmy would complain about a poor student scraping a web page in order to learn something.
Seeking distinctions is pretense. They’re just shuffling cards.
You can ask about models made from public-domain data, and most critics will not budge an inch. Mentioning copyright is working backwards from a gut feeling. The ones who say, sure, okay, it’d be different if– - maybe they have a consistent rationale. But even some of them haven’t examined how they’d feel about this technology, if all their complaints were addressed.
Filesharing is very different from profiting off of your mass scale piracy.
Correct - only the filesharing is against the law. Training is transformative use.
You can’t cram a billion images into one gigabyte. They’d be one byte each. What these models do is very different from the bootlegging you’re trying to make it sound like.
“Lemmy really hates it when a corporation murders people, yet they never say anything about individuals who kill in legitimate defense of themselves”
That’s pretty much how you sound, buddy.
If the AI created was really useful and made open to all freely, a different song would be sung, you know.
Right now, it is simply as useful as putting gas on a fire to extinguish it.
Arguments are easy when you make shit up.
Anyone who’s not frothing mad about spicy autocomplete sounds like a murderer, says the shrillest kneejerk response in a competitive field.
Thinking it’s only about the LLMs is pretty reductive, though I guess it makes for an easily digestible statement.
We don’t ask people to be mad about the tool, it’s about how we’re being told The Tool is the most important human endeavor, while the mass usage of The Tool has yet to prove more worthwhile than having a chat with a friend, all the while we create a lot of ethical, ecological and economical debt in order to keep evolving The Tool, and funneling most of the products of that debt into the hands of people who already have more than enough.
So The Tool as of yet has had way more negative direct and side effects, by the combined effect of its (mis)use, its Herculean development and the massive, global pitch sale effort invested in deploying it to ubiquity, than any projected positive outcomes that comes out of the mass deployment and usage of The Tool.
It has its very specialized used. It should have remained in the labs and backends where it belongs.
This is what - to me, one of the Luddite who would rather stop this cancer from growing - we are mad about, and why, given the scale of things at this moment and what is projected, think people should be more in the know of the revolting realities that are not said in all those nice press releases and consumer expos.
