Inside the Warehouse Where Amazon Scans and Destroys Books For AI Training
- Reference: 0185198086
- News link: https://news.slashdot.org/story/26/08/26/1857240/inside-the-warehouse-where-amazon-scans-and-destroys-books-for-ai-training
- Source link:
> [...] What does it look like when they cut the spines off the books?
> It's a machine people operate. Each of these machines is just like a little workstation, and there'll be one person at each of these stations. It's really safe because it has a little cover and it has a little area where you slide the book into, remove your fingers from the area, and then you press a button, and it just comes down and slices it. And then you remove the books after the blade is gone.
>
> After they cut the spines they have these kind of library carts, kind of like how you would stack books, but instead, it's a stacks of paper and little cardboard things to separate -- I'm guessing -- the different books. I saw the pages being thrown in a shuttle after they were scanned. We were walking by and we went over to one of the shuttles and looked into it because we could tell they were throwing the old books in it or the cut books in it, and it's just a bunch of loose paper like just sheets thrown in there.
>
> What do you think about Amazon doing this?
> At first what they were telling people is that it was for Kindles, but I just didn't believe that. I instantly was like, I don't think that's how they do it because of publication rights and stuff, copyright and everything. Then I saw it was for AI. I don't like it only because I wish I could take these books. There's so much knowledge and so much stuff in them and some of them look like they might be rare, and I've heard that they order rare books, and that's why I say some of these are so obscure. I definitely don't like the idea that they can't be reused or anything like that afterwards. They're reducing the amount of available copies for other people.
[1] https://news.slashdot.org/story/26/08/17/1644216/tracking-rare-books-leads-to-an-amazon-ai-training-facility
[2] https://www.404media.co/inside-the-warehouse-where-amazon-scans-and-destroys-books-for-ai-training/
[3] https://slashdot.org/~alternative_right
Technology to the rescue again (Score:1)
What a great thing technology is. Instead of burning books like in the olden days to suppress information we can now scan and destroy them and lock the information away behind so-called AI. The paper can then be recycled into documents that we must all carry around to prove we are good little humans... papers please.
Re: (Score:2)
> What a great thing technology is. Instead of burning books like in the olden days to suppress information we can now scan and destroy them and lock the information away behind so-called AI. The paper can then be recycled into documents that we must all carry around to prove we are good little humans... papers please.
This is copyright law issue, not a technology one.
Once you scan a book, you have to destroy the book.
Google books had page turning robots at libraries and scanning books 10-20 years ago.
Full circle (Score:2)
Amazon went from selling newly printed books to buying books and throwing them into the shredder.
Jeeze (Score:2)
Where do people get off thinking they can tell people what they can do with their property? If someone wants to buy a Picasso and wipe their ass with it, it's none of my business. I am not entitled to other people's property. I'm thinking AI delusion may be more powerful than TDS... people lose all sense when they get so emotional about these subjects.
Re: Jeeze (Score:2)
You have lost all sense when you are arguing no one should be upset if you wipe your ass with a Picasso. It's a ridiculous example, but you are clearly a ridiculous person
In a sane world... (Score:2)
...the scans could be donated to free online libraries.
Sadly, the copyright owners would sue, even though they aren't available for sale any more.
If itâ(TM)s okay for Amazon⦠(Score:2)
If itâ(TM)s okay for Amazon to scan and digitize thousands or millions of books, then itâ(TM)s okay for you to do it too. Just a thought.
Gen X might remember... (Score:2)
The headline reminds me of "Johnny 5 need input"....from Short Circuit series of movies
The only books they're destroying... (Score:5, Informative)
....are books that they can buy by the meter, for pennies on the dollar, if that much. These are surplus books that publishers - the world's leading destroyers of books - would have to destroy anyway.
Trainers can't afford to buy actually rare, desirable books.
It's a shame that all of this data is locked up on private servers, but that's really more of a problem of the copyright system than anything else. Anyone who scanned these books would be banned from making it public.
Re: The only books they're destroying... (Score:2)
Yeah, Amazon is famously hurting for cash.
Re: (Score:1)
UDHR27 makes it clear that these technological advancements are for all of us, free.
copyright and patents are an illegal monopoly on culture and technology.
Re:The only books they're destroying... (Score:4, Informative)
No one is destroying rare books to scan them for AI. They are scanning books that are cheap and available at bulk rates: books sold by weight on pallets.
I am a book dealer. I have run an online book business, I currently own a small bookstore.
I have personally sent entire shipping containers of books to be recycled. There is just that much printed materiel that no one wants.
Re: (Score:2)
I mentioned this to the other guy but yeah they were destroying rare books. A rare book seller caught them by sneaking an Apple Air tag into one of his books that he shipped to them. That's the only reason we know about this.
Both things can be true. There can be a ton of books that nobody wants and to get thrown out and Amazon can be buying those books in addition to Amazon buying rare books to destroy for their AI model. Because of course they're going to want rare books if they have rare books in thei
Re: (Score:2)
> I mentioned this to the other guy but yeah they were destroying rare books. A rare book seller caught them by sneaking an Apple Air tag into one of his books that he shipped to them.
There's rare and then there's "rare." From what I've seen, the people involved have declined to say what the books involved were. About the only reason I can think of to hide that is because they know that people won't be as outraged when they find out what the "rare" books actually are. However, it was a bulk lot of something like 1,000 books, so it's a pretty safe bet that they weren't exactly valuable.
The 1973 manual for a toaster that I found in my grandfather's attic is certainly rare, but we're not lo
Re: (Score:2)
> I mentioned this to the other guy but yeah they were destroying rare books. A rare book seller caught them by sneaking an Apple Air tag into one of his books that he shipped to them.
No. They are not.
If you had bothered to read the article you reference, you would find that the researchers paid to have trackers inserted into a shipment of books that a book dealer sent to Amazon to be scanned so they could track where the books actually physically went once Amazon took the shipment. There were no "rare books" involved. There was only more FUD about the potential for rare books to be destroyed. The actual books sent were all modern prints of books identifiable by ISBN.
It wasn't researchers (Score:2)
It was a rare book dealer. What's probably happening here is there's a group of researchers who did the same thing that the rare book dealer did. Again both things can be true at the same time.
If you would bother to Google The articles involved you would know that. But like a lot of people around here I'm sure you're Google is broken. Either that or you're using AI to Google instead when which case yeah you don't know what you're getting...
Citation please [Re:The only books they're des...] (Score:2)
> I mentioned this to the other guy but yeah they were destroying rare books. A rare book seller caught them by sneaking an Apple Air tag into one of his books that he shipped to them. That's the only reason we know about this.
Citation, please.
Re: The only books they're destroying... (Score:2)
No it seems like lots of the orders are [1] very particular [rnz.co.nz] or maybe automated based on price or some other factor. Books in linked article are not exactly valuable but pretty unique and potentially not many copies around.
[1] https://www.rnz.co.nz/life/books/wellington-second-hand-book-store-s-mysterious-orders
Re: (Score:2)
That was an interesting article. Thank you for linking it.
None of the books referenced were rare books. Many were of limited marketability due to the subject matter or language -making them of interest for AI training- but still not rare books:
> "The buyer wanted books with ISBNs (a unique code used to identify a specific book, edition, or format) from the 1970s and 1980s, she told RNZ’s Midday Report."
> "They were local non-fiction niche books, so New Zealand local history, a tourist pamphlet on Waito
Re: (Score:2)
> That was an interesting article. Thank you for linking it. None of the books referenced were rare books.
How do you know that?
Nothing in the text you quote stated or implied "these are not rare books".
Many of them seem to be items of local interest only, or items of ephemeral interest only; it is very possible that these would be rate. In fact, the quoted text said "niche books", and these would be rare pretty much by definition: if they were widely circulated, they wouldn't be niche.
> "The buyer wanted books with ISBNs (a unique code used to identify a specific book, edition, or format) from the 1970s and 198
Re: The only books they're destroying... (Score:2)
They are taking books out of circulation for human use, and making them into effectively an AI information monopoly.
Imagine a future where if you want to know something you have to ask AI. You can't even look it up yourself because those resources were pulped ages ago.
Re: (Score:2)
The point is that these are books already being taken out of circulation... these are pallets of unsold books that nobody wanted, otherwise destined for the recycling center. They're just inserting an extra stop at the scanner before going on to the recycling center.
I don't support LLMs destroying lots of things, but the problem part of this is not "destroying books slated to be destroyed".
Re: The only books they're destroying... (Score:2)
Reminds of Jean Baudrillard, the philosopher who predicted representations of reality replacing reality, leaving the *desert of the real* (source of The Matrix quote) behind. Except it has reached a new level, where more abstract representations are replacing lesser ones:
* Simulation is the situation created by any system of signs when it becomes sophisticated enough, autonomous enough, to abolish its own referent and to replace it with itself *
That's only true because they got caught (Score:1)
So they want as much data as they can get and they were actively seeking out rare books. They had gone to a rare book seller and they were buying them to destroy them. Because of course they want rare books if they have rare books in their model their model will be better than somebody who doesn't have rare books.
The way we found out they were destroying rare books is that the rare books seller got a huge order and he got suspicious so he snuck an Apple Air tag into one of the books and monitored where
Re: (Score:1)
Fahrenheit 451
Citation needed? (Score:2)
Rei: Do you actually work for Amazon in a capacity where you're informed on how they're obtaining these books?
I'm asking for a couple reasons.
1. We all know Amazon has plenty of money to be able to afford to buy rare, out-of-print books that are difficult to obtain. I'd be surprised if they refrained from doing so if that meant their AI risked not having some of that knowledge/information while a competitor's did. I'm not saying the guy tasked with being manager over the disposable labor doing the book sca
Re: (Score:2)
> Trainers can't afford to buy actually rare, desirable books.
Note that "rare" and "desirable" are different things.
You seem to be implying that anything that's rare would be desirable, but that is not always the case.