Latest / The WritersWeekly Podcast / EPISODE 51: AI ABOMINATION! AI Companies are Buying Millions of Rare Books, Scanning Them, and Then SHREDDING THEM!!!

Transcript

EPISODE 51: AI ABOMINATION! AI Companies are Buying Millions of Rare Books, Scanning Them, and Then SHREDDING THEM!!! · The WritersWeekly Podcast

0:52
Brian: But for now, let's get on to this episode of the Writers Weekly Podcast.
0:57
Brian: Hello again, and welcome to another episode of the Writers Weekly Podcast.
1:06
Brian: I'm your host, Brian Witten, managing editor of Writers Weekly, and with me, as always, is Angela Hoy, owner operator of BookLocker.com and CEO plus.
1:18
Brian: He is the original originator of Writers Weekly.
1:23
Angela: I'm pretty uh I always said a bad word.
1:25
Angela: I know.
1:26
Angela: I'm very upset today.
1:27
Brian: Well, we're gonna we're gonna be talking about a uh subject that has recently uh hit the I guess hit the public view.
1:35
Brian: It's got a lot of people's panties in a wad, and a pair of those panties belong to Angie.
1:40
Brian: So um I think she's gonna take off on this.
1:44
Brian: Take it away.
1:46
Angela: Oh, good.
1:46
Angela: Give me the mouse.
1:47
Angela: I'm in control.
1:48
Angela: Yeah.
1:49
Angela: Um this episode's called AI Abomination.
1:53
Angela: AI companies are buying rare books by the palette, ripping them apart, scanning them, and then shredding them.
1:59
Angela: I was talking on the phone last night to Professor Clayton Jones, everybody's favorite editor and book marketer, and he he made a very good point.
2:07
Angela: He said, Imagine museums selling off all their artifacts knowing they will be destroyed.
2:12
Angela: That's exactly what's happening with old rare out-of-print books right now.
2:16
Angela: Booksellers are selling them to AI companies knowing they're going to be destroyed.
2:22
Brian: So, um one of the big big issues when AI first got started was, you know, they were they were scraping the internet and taking information, including uh scanning online books to train their AI models.
2:38
Brian: And of course, that caused uh a lot of consternation with many people.
2:45
Brian: Um and the issue was addressed, and I guess there have been uh a few decisions made that that that's a copyright infringement for AI companies to just go in and start uh peeling out you know people's information in their books and just plugging it into their into their programs to teach AI.
3:07
Brian: Right.
3:07
Brian: So they found themselves a little workaround here.
3:10
Brian: Angie, why don't you explain it?
3:12
Angela: Yeah.
3:12
Angela: Well, first of all, if any of you have any rare or out-of-print books on your bookshelf, keep them.
3:18
Angela: You may end up owning the very last copy of that book on earth.
3:21
Angela: That's what's happening.
3:22
Angela: Millions and millions of books are being destroyed.
3:24
Angela: Uh before we get into this shocking story, you need to know that the largest player in this mess is anthropic, which recently settled the largest copyright infringement lawsuit in history.
3:34
Angela: That just happened a few months ago.
3:36
Angela: But they make so much money taking information created by others that the one point, I think it was one point two billion dollar settlement was likely well worth it to them.
3:46
Angela: They're making they're making you know billions, trillions of dollars.
3:49
Angela: Yeah.
3:50
Brian: So like I explained uh a minute ago here, so large language models, and that's known as LLMs, uh, that power things like ChatGPT and other similar AI systems, uh, have in the past uh been primarily trained using information from the internet, so digital uh stuff.
4:09
Brian: Yeah, you know, articles, uh human interaction online, and ebooks.
4:15
Brian: But uh today almost all of the information available to mankind has already been scraped up from the internet.
4:22
Angela: Yeah, they've run out of data.
4:23
Angela: Yeah, yeah.
4:24
Brian: Um AI companies are now hungry for more and more data, and they're coming after um out-of-print book, uh out-of-print and rare books, yeah.
4:35
Brian: Uh actual physical books.
4:36
Brian: Anthropics lawsuit only dealt with ebooks, and they had obtained pirated copies of these e-books and fed that text uh from the e-books into their AI systems for training.
4:49
Brian: They didn't purchase those books, and their actions were definitely uh copyright infringement.
4:54
Angela: Right, and we'll get more into that in a little bit.
4:56
Angela: Uh what we're talking about today is completely different from the ebook situation, which was egregious on its own.
5:03
Angela: It's what they've now been caught doing print books, and they were very secretive about it, but uh somebody got a hold of some internal documents, and you would not we're gonna give you a couple quotes from those today, and you're not gonna believe what Anthropic was doing.
5:17
Angela: So shock waves are emanating through the author, publishing and book loving communities this week.
5:22
Angela: I'm I'm seeing more and more posts.
5:24
Angela: I'm seeing uh videos on TikTok of what's happening, um, the machines they're using.
5:30
Angela: People are putting videos of those machines on TikTok.
5:33
Angela: And a stunning and downright nauseating report from 404media.co has revealed that AI companies are ordering old and rare books, thousands and even up to a million different titles at a time in one order.
5:47
Angela: So once they receive those books, they put those through these like high-speed machine, this high-speed machine that cuts the spines off.
5:55
Angela: I saw a video of that today.
5:56
Angela: It's called a book debinder.
5:58
Angela: And those were originally recycling and and destruction equipment, and they were designed to slice the glue spines off.
6:07
Angela: So so basically it's like an old paper cutter, except it's a big industrial thing, and it slaps down on the um spine of the book and just cuts off.
6:14
Angela: All the glue and stuff, right?
6:16
Angela: But the paper then the pages are all separate and you can scan all those pages.
6:19
Angela: Yeah.
6:20
Brian: And and originally that was used to to like recycle paper from old books and stuff like that.
6:27
Angela: Yeah.
6:27
Angela: And they can cut off the glued spines from covers at speeds exceeding 1200 books per hour.
6:34
Angela: And so back in the day it was you know, a publisher printed way too many books, they weren't gonna sell, and they would hire these companies to cut cut up the books so they can never be sold and then recycle the paper.
6:48
Angela: That's basically what those were designed for.
6:50
Angela: Now they're being used for something completely different.
6:52
Angela: And then the AI companies use industrial grade imaging equipment.
6:57
Angela: They ski that they use those to scan the pages and feed them into their systems to train the AI systems.
7:02
Brian: Yeah.
7:03
Brian: Um, Futurism accurately described this process uh when they wrote um to those buildings uh to those building large language models, books are nothing more than uh fodder to be devoured and mass before being spit out like fishbone.
7:21
Angela: That's exactly what's happening.
7:22
Brian: Um according to ISBNDB.co, um while the litigation is still unfolding, Anthropic has already been quietly building an escape hatch, a procurement operation so unusual and so methodical and so deliberately low profile that it reads more like a Cold War logistics program than a Silicon Valley product strategy, and they call it Project Panama.
7:49
Brian: And that's according to ISBNDB.co.
7:53
Angela: Yeah, we're we're quoting our and we've gone we're going to put links to all these sources underneath this episode at writersweekly.com slash podcast, episode 51.
8:01
Angela: So, listeners, do you agree with what AI companies are doing?
8:06
Angela: Literally, they went from scraping the internet to scraping bookstores and booksellers and book warehouses.
8:12
Angela: And and once they scan those books and just they're gone.
8:16
Angela: Yeah.
8:16
Angela: Nobody can ever have a copy of that book again.
8:18
Angela: And sadly, there's no shortage of booksellers willing to sell these rare and out-of-print books to AI companies, knowing what's going to happen to them.
8:26
Angela: And it's all about the Benjamins, right?
8:28
Brian: Do do we know what they're doing with the pages once they get done scanning these?
8:32
Brian: I mean, are they are they burning them, shredding them?
8:34
Brian: They're either shredding them or pulping them.
8:35
Brian: Pulping them, yeah.
8:36
Brian: Okay.
8:37
Brian: Yeah.
8:38
Angela: And uh one of those companies that is willingly selling to the AI companies, again, pallets and pallets of books is isbndb.co.
8:45
Angela: Oh, or is it com?
8:48
Brian: Well, we're gonna have to clarify that after the Well, you know what?
8:51
Angela: We'll um you know what you you uh you talk for a bit and I'll look it up.
8:55
Brian: So my fault and ISBN DB stands for ISBN Database.
9:02
Brian: That company claims to have uh the world's largest book database, and it's now offering bulk buying options to AI companies of up to one million titles per order, including quote, older, rare, and specialist volumes, unquote.
9:21
Brian: Uh, according to that company, and we'll include a link under this episode at writersweekly.com podcast, uh actually slash podcast, and this is episode 51.
9:33
Brian: Um, quote, in 2025, a federal judge, William Al Su handed down one of the most consequential copyright rulings in AI history.
9:44
Brian: Anthropic, they also uh have created clawed AI, yeah, had purchased and physically destroyed over two million books as part of a strategy to acquire AI training data.
9:57
Brian: That judge said that they were not violating copyright law.
10:02
Brian: Uh Anthropic's large-scale version of the practice was performed under an internal effort known as, again, Project Panama.
10:12
Angela: We're gonna mention that a few times today.
10:13
Angela: Okay, ISBN DB.co redirects to.com.
10:17
Angela: So either one will work.
10:18
Angela: Okay.
10:19
Angela: I looked it up.
10:20
Angela: Yeah.
10:20
Angela: Okay, I turned my phone off again.
10:22
Angela: Now, um, tomshardware.com, which has nothing to do with uh, it's not a hardware store.
10:27
Angela: They said ISBN DB, an online database that reportedly has over 111 million cataloged books, has been a long favorite platform for booksellers, libraries, and distributors to sell books.
10:41
Angela: With the explosion of the AI industry, ISBN DB has pivoted its business to offer specialized services to bulk purchase books, to bulk purchase books.
10:50
Angela: They purchase them and then they sell them to the AI companies.
10:54
Angela: So they're like an intermediary intermediary between somebody who's selling books and then they they get all these big orders together for for the AI companies.
11:02
Angela: Yeah.
11:03
Angela: Which is interesting because we're quoting their website today.
11:06
Angela: And they don't seem to look at it very favorably in their writing about the practice, but they're they're needing they're making a lot of money from it, yeah.
11:13
Brian: Um one small bookseller uh who was interviewed, he said that he went from selling from around 20 books a week to several hundred, again, that's each week.
11:24
Angela: Yeah.
11:25
Angela: And I read a quote from him, and he's not happy about the practice, but you know, he needs the money.
11:29
Brian: But he needs the money.
11:30
Brian: Right.
11:30
Brian: So he'll do it.
11:31
Brian: Yeah.
11:32
Angela: So we disagree with the federal judge who made that ruling.
11:35
Angela: I I mean, just it's ridiculous.
11:37
Angela: Now we're going to quote the doctrine that he was going by in just a minute, but here is the typical language you see on the copyright page of a book, and I'm going to emphasize some of these words.
11:51
Angela: No part of this publication may be reproduced, stored in a retrieval system.
11:56
Angela: Sounds like AI doesn't, or transmitted in any form or by any means electronic, mechanical, recording or otherwise, without the prior written permission of the author.
12:07
Brian: Okay, however, this this judge ruled that it was fair use under the first sale doctrine, and that's section one zero nine of the copyright act.
12:17
Brian: The first sale doctrine written way back in 1908.
12:22
Brian: Nineteen oh eight before we had computers, before we had AI, before we had any of this stuff, holds that once a copyright holder has authorized the initial sale of a physical copy, their control over that specific object ends.
12:38
Brian: So if you buy a book, then you own the book, you can read it, sell it, lend it, etc., and the author has no further claim on what you do with that physical item.
12:47
Angela: Right.
12:47
Angela: And I agree with that.
12:49
Angela: But they're doing II companies are doing something, you know, completely different.
12:53
Angela: Yeah.
12:53
Angela: However, that doctrine is 118 years old.
12:58
Angela: Yes, I pulled out my spreadsheet and figured that out.
13:01
Angela: It long predates ebooks, it predates scanning equipment, it even predates the internet, you know, by predates copy machines.
13:10
Angela: Right, yeah.
13:11
Angela: Exactly.
13:12
Angela: Yeah.
13:12
Angela: The doctrine does not state that the owner can make new copies of the book, including electronic copies.
13:18
Angela: And in scanning a book, whether you whether you keep the text together or you disseminate it in a variety of ways through an AI system, you're still putting copies out there for the masses.
13:27
Angela: Yeah.
13:27
Angela: And rewriting a copyrighted work is still a violation of that copyright.
13:31
Angela: And when AI spreads the book's information across the globe, even changing the meaning of the text, that's still copying that copyrighted work.
13:38
Brian: Yeah.
13:39
Angela: I believe that soon that doctrine and copyright law will be updated to address today's technology.
13:45
Angela: And at the end of this episode, we're going to share with our listeners how they can help make that happen, and it's very easy.
13:51
Brian: This is uh this is interesting.
13:53
Brian: AI companies are specifically looking for books uh published prior to 2022, and of course, books that are far older than that.
14:03
Angela: That's interesting.
14:04
Brian: Um, their reasoning is that books published prior to 2022 weren't written using AI.
14:10
Brian: Um it's important to know that one problem AI has now is that it's regurgitating its own information.
14:16
Brian: We've talked about this before.
14:18
Brian: It's basically copies of copies of copies.
14:20
Angela: Right.
14:21
Brian: And it's getting tainted by incorrect information.
14:24
Brian: Uh the AI systems can't keep track of what's already out there, and they can no longer keep track of what's true and what's fake.
14:31
Angela: Right.
14:32
Brian: Uh so did you know that some authors are purposely putting out garbage books, specifically um intent on filling the AI systems with gobbledygook, and there's no law against that.
14:45
Angela: You know, if I'd if I'd read that authors were doing that, but they hadn't mentioned AI, I would have gone, oh, that's just awful.
14:51
Angela: But yeah, there are authors that are purposely trying to fool the AI systems to get back at them for taking away from our livelihood.
14:59
Angela: Yeah.
15:00
Angela: Uh along those lines, um, again, according to isbn.com or.co, they both work.
15:08
Angela: Uh their article includes this text.
15:12
Angela: This is interesting.
15:13
Angela: A $1.5 billion lesson in data hygiene.
15:17
Angela: That's a great term, data hygiene.
15:19
Angela: Yeah.
15:20
Angela: For the better part of a decade, I'm quoting again, the AI industry operated under an open secret.
15:26
Angela: Training data sets were built from content nobody had paid for.
15:30
Angela: This is where this is what got anthropic into trouble.
15:33
Angela: LibGen, Books 3, Z Library, and the Pile were the pipelines through which hundreds of billions of words flowed into the weights of the world's most powerful language models.
15:44
Angela: Now, I knew about LibGen, and I've heard of a couple of the other ones, but I've actually looked at LibGen when the anthropic lawsuit first came out.
15:51
Angela: LibGen is library genesis.
15:54
Angela: There is no known official corporate owner, legal entity, or known individual proprietor.
16:00
Angela: Why?
16:00
Brian: Yeah, it's because they're able to hide it that well.
16:03
Angela: Well, yeah, but they're they're running a completely illegal operation.
16:06
Angela: Yeah.
16:06
Angela: It operates as an anonymous, decentralized, quote, shadow library founded around 2008 by hidden contributors who uh many people suspect they're located in Russia.
16:16
Angela: And it is fed through community donations and mirror networks.
16:22
Angela: Brian, you know what a community donation is?
16:25
Brian: That sounds like dark money to me.
16:27
Angela: All right, right.
16:28
Angela: Uh the ebook pirating companies allow people who've purchased ebooks to upload those ebooks to their systems so that other people can download the illegal pirated copies of those ebooks for free.
16:38
Angela: And uh there's copyright laws in pretty much every country.
16:45
Angela: Yeah.
16:45
Angela: Yeah.
16:46
Brian: Sound sounds a lot like what was going on with um Pirate Bay.
16:50
Brian: Oh, right.
16:51
Brian: If you owned a copy of uh, you know, of music or a video game or anything, you could upload it uh using torrent software to Pirate Bay, and then anybody who wanted could just download it.
17:04
Brian: Yeah, you know, the way it was.
17:06
Angela: Right.
17:07
Angela: Um Anthropic and other AI companies began to download massive numbers of illegal copies of those books to train their systems, and then they ran out.
17:15
Angela: Anthropic's recent one point two billion dollar lawsuit was just the beginning, and uh more lawsuits have already been filed against other AI companies for doing the same thing.
17:25
Brian: Now, how many illegal copies of ebooks did Anthropic obtain?
17:30
Angela: Everybody guess that before Brian gives you the answer.
17:35
Brian: Reports state that there are over seven million.
17:39
Angela: Yep, downloaded seven million illegal copies of ebooks.
17:43
Angela: And if you're listening to this and you're an author, your book may have been involved in that as well.
17:48
Angela: Yeah.
17:49
Brian: Uh Anthropic then partnered with used book wholesalers, and they started buying print books by the palette.
17:56
Brian: So millions of books across all genres, and they've they've all been, they've got their their uh spines cut, they're torn apart, they're scanned, and then they're destroyed.
18:11
Brian: Yep, and all that's left is pulp.
18:12
Brian: That's it.
18:13
Brian: Yep.
18:14
Brian: And um, this is reminding lots of folks about George Orwell's book 1984, where the government burned books and then changed historical records to control the official narrative.
18:26
Brian: In the end, only the revised text remained, and that's that's what could be happening now.
18:32
Brian: You've got these AI companies, they're the ones that control what the output is on their AI if they want to.
18:38
Brian: Right.
18:38
Brian: So they can they can take any any inconvenient little tidbits of history and simply tell the IAI don't don't put that out there and say this instead.
18:48
Angela: Uh this gentleman contacted me a few months ago.
18:51
Angela: He wrote an article for us, not on this topic, and he was using AI just to test the system, and he said it was misquoting the Bible.
18:58
Angela: Yeah.
19:00
Angela: I wonder how many Bibles have been chopped up and cost.
19:04
Brian: I know I know that China um, you know, published its own version of the Bible, basically singing the praises of the Chinese government.
19:12
Brian: Right.
19:12
Brian: Yeah.
19:13
Brian: You know, and and um so I can see AI companies doing the same thing.
19:17
Brian: If you don't like what the Bible says, you just have it say something else now when people search for it.
19:23
Angela: Right.
19:24
Angela: Uh back to TomsHardware.com.
19:26
Angela: Um, it's they said AI companies seemingly have a new target in their sites.
19:32
Angela: Human hum humanity's literary history.
19:34
Angela: That's yeah, that's what we were just talking about.
19:37
Angela: They're gonna change everything.
19:38
Angela: Yeah.
19:38
Angela: So you mentioned before other lawsuits, Tom's Hardware also wrote a coalition of publishers recently filed a lawsuit against Google.
19:47
Angela: We knew about that one, accusing the tech giant of allegedly and illegally using millions of copyrighted books to develop its Gemini AI models.
19:54
Angela: Now, Google got sued several years ago for the same thing.
19:59
Angela: They were scanning books, but they were borrowing books from libraries and scanning them, and they were buying books and scanning them, and a judge ruled that was fair use.
20:07
Angela: Yeah.
20:09
Angela: So anyway, Google got caught allegedly downloading illegal copies of e-books and doing the same thing that Anthropic did.
20:16
Angela: MATA, our least favorite company in the world, the owner of Facebook and Instagram, has also previously been sued in U.S.
20:23
Angela: courts for downloading millions of pirated books.
20:25
Angela: They're all in this big race.
20:27
Angela: Yeah.
20:27
Angela: You know, to be the first person to have every bit of information ever created in the history of mankind in their system.
20:35
Brian: And I saw an article, I think, just a couple days ago, that Mark Zuckerberg is whining and crying because Anthropic and this other companies got far more AI than he does, and he thinks that they ought to be forced to just share release all of their all of their proprietary stuff so that he can have some too.
20:53
Brian: Yeah.
20:54
Angela: I we will put a link to that article under this episode.
20:57
Angela: Yeah, we put it in the case.
20:58
Brian: It's like the kid in the sandbox that's all pissed off that the other kid's got the Tonka truck and he wants it now.
21:02
Brian: It's just it's it's crazy.
21:05
Brian: So uh David Sachs, who is a venture capitalist and he's the chair of the President's Council on of advisors on science and technology, there's a mouth.
21:16
Angela: That that is a that's a big title, yeah.
21:18
Brian: Um wrote on a wrote in a post on X that Anthropic maintains that it is entitled to train for free on all the world's output, even if the author objects.
21:30
Brian: But if a competitor trains on anthropic's output after paying for it, that's IP theft.
21:37
Brian: Uh the hypocrisy is just is breathtaking.
21:40
Angela: Yeah, they're saying we can take everything from normal folks like you and me and use it for our systems to make a bunch of money.
21:46
Angela: But if anybody uses our system to do the same thing, oh yeah, that's that's not fair.
21:52
Angela: Yeah.
21:52
Angela: That's a lot of hypocrisy there.
21:54
Angela: Remember, and we've covered MATA a lot, and we'll put a link to some of those articles.
21:59
Angela: articles under this episode, um, made a you know, putting scammers over their own users, earning billions and billions of dollars from scammers, um uh feeding content to children that led to some children's suicides.
22:13
Angela: They recently lost a huge lawsuit and they actually last week uh they lost the appeal on that and I can't remember how much money they owe that one person one girl it's like an eight million dollar three million three million it was them and YouTube and two other companies two other social media companies had already she was suing based on the fact that they had made their algorithms addictive right yeah and it dealt with uh body image and things like that but they they made a apparently according to internal documents obtained by Reuters knew that their system was feeding children these harmful things and they they they first stopped it but then they realized their their profits were diving so they started it back up again.
22:54
Angela: And now there are hundreds of lawsuits pending against them for the same thing.
22:59
Angela: So MADA in their internal documents they only banked on getting sued in the United States yeah they actually in the documents is you know they say well it'd be worth it to have be fine this much money because we made this much money right what what they didn't think was what wasn't in the numbers was gee what if we get sued in other countries and that's happening too right a bunch of other companies countries suing well maybe anthropic made the same assumption copyright laws vary from one country to another I mean in the European Union their digital single market directive allows rights holders to opt out of text and data mining why isn't the United States there yet how many books copyright copyrighted in UK countries has anthropic and other AI companies used to train their systems I imagine a lot.
23:50
Brian: I wonder if there's any way to find out if they're just buying up pallets of books and chopping them all up.
23:55
Angela: I mean I'm sure a lot of those books were printed published in other countries and subject to other countries' copyright laws.
24:02
Brian: So now Anthropic and all the other AI companies have to worry about what made us worrying about uh oh you know we've had this one here's one lawsuit uh oh here's another one another one another one and it's not just going to be individuals it's probably going to be governments yeah assuming as well so according to an Australian website news.com.au um Australian book dealers are urging Anthony Al Albanese to protect local titles from the practice under the prime minister's newly announced office for AI um office for AI yeah um U.S.
24:40
Brian: copyright law and its definition of fair use is different from that in Australia.
24:45
Brian: So now Australia's coming after you've already got these other other countries starting to to maneuver into place to to fight this.
24:52
Brian: Right.
24:53
Brian: In the article on uh news.com.au uh they quote Dawn Albinger Albinger yeah I think so who is the owner of Queenland's largest secondhand bookstore Archives Fine Books and she said I'm quoting when content is extracted from books to train AI models it raises very significant ethical and legal concerns.
25:18
Brian: It also strips away the structure context and meaning intended by the author undermining the integrity of the creative work and reducing the viability of a healthy book industry.
25:32
Angela: Nobody could have said it better.
25:33
Angela: Yeah yep now the Australian Booksellers Association chief executive Susanna Bowen she said almost the exact same thing in an article I read and according to news.com.au rare booksellers across Europe are seeing a similar surge in purchase requests which they suspect are coming from AI labs in America.
25:52
Brian: They know where they're coming from yeah uh AI companies clearly know that they're doing something wrong because they're making all of these purchases on the down low.
26:02
Brian: Yep they're not there you know yeah why not just tell everybody yep this is what we're doing.
26:06
Brian: Yeah right buying millions and millions of books going to scan them and chop them up wait till y'all hear what they actually said for example um Dutch news site uh BNR reported that Peter DeVries and uh antiquarian bookseller in in Harlem Har Harloom I'm sorry the pronunciation's different in in in uh the Netherlands yeah um which is okay so Harloom it might be Harlem I don't know anyway well anyway uh based in the Netherlands Netherlands um recently received an email from a person calling herself Natalie from Singapore and she wrote that her company was starting a new project focused on collecting books in multiple languages currently mainly in English Natalie Natalie how many people in Singapore name Natalie I'm sure it's plenty.
27:05
Brian: I I to be honest I don't know what a Singaporean name normally sounds like also it's an interesting spelling of Natalie.
27:13
Angela: Yeah anyway the email said and I quote we have compiled a very extensive list of editions that we are currently trying to acquire and we plan to place a fairly large order.
27:23
Angela: Now the email attachment contained a list of 3,000 English language titles they were organized by ISBN and ranged from book books on fairy tales to fork folklore technical manuals and even science text so some of this might just be an ISBN number you know they don't even care what the what the subject matter is.
27:45
Angela: They're just plucking up as much as they can but it looks like they're it's not just books they're getting scientific papers they're they're they're getting as much as they possibly can and again it's all going in the trash now booksellers they're catching on they know these are coming from AI companies but they're still doing I I don't know how well first of all there's the financial aspect yeah and but eventually they're gonna run out of books to sell.
28:10
Brian: Well true yeah true um the NL Times uh reports that books booksellers in the Netherlands Switzerland Spain and Germany uh have received nearly identical requests so this person's basically copying pasting and asking everybody for their books if the AI companies think everything is above board um why are they being so secretive right and why have they kept this practice so hush hush and why are they using third parties to purchase palettes of the of the books uh to the tune of millions going out and buying them themselves yeah now this is going to shock all of you like it shocked me I'm still nauseated about it listen to this okay according to the Dallas Express an internal planning document obtained from Anthropic stated that and I quote Project Panama is our effort to destructively scan all the books in the world you should probably say that again that's this this is from Anthropic's own internal documents.
29:13
Angela: Project Panama is our effort to destructively scan all the books in the world destructively scan that means scan and scan and junk yeah right the same document added and I quote we use a soft code name for it because we don't want it to be known that we are working on this this document is visible to all anthropic employees but you should avoid talking about it in public areas and the fact that we are working on this should not be shared with anyone outside Anthropic.
29:39
Brian: Well too late now well then they got sued and all this stuff it came to light so yeah um yahoo.com posted a great article and we'll post a link to that um and all of the other resources on writersweekly.com slash podcast and this is again episode 51 and they wrote that uh quote a used bookseller gets an order for 400 volumes history botany regional law german language economics the subjects share nothing in common except for one detail every title has an ISBN there it is again that's the fingerprint AI companies uh or their intermediaries shielded by non-disclosure agreements also called NDAs are specifically buying physical books from secondhand markets slicing off their spines for high speed scanning and then pulping the originals.
30:29
Angela: Yeah and then they go on to say the reason is brutal brutally simple the internet is so saturated with AI generated text that pre-2022 printed books represent the la last large uncontaminated reservoir of human authored knowledge a dynamic that reflects the same appetite for data driving ventures like the Stargate project with what's that okay so uh project Stargate uh that's a $500 billion dollar artificial intelligence infrastructure joint venture announced in 2025 um it's a private sector partnership backed by OpenAI SoftBank Oracle and MGX anthropic calls its AI model Claude uh in courtroom documents obtained by the Washington Post Anthropic's project Panama used contractor datamation to scan books for Claude's training data the Washington Post also talked about these huge orders being generated uh they wrote that booksellers know it's an AI buyer because of each other's abnormal or each order's abnormal volume subject agnostic orders and total indifference to pricing subject agnostic I like that yeah they don't they're just getting whatever they can ordering you know 10,000 copies of a specific topic right you know and different books covering one thing it's just ordered everything they can get their hands on.
32:00
Angela: Yeah so you said total indifference to pricing I I read in other places online as well in articles AI companies aren't negotiating prices.
32:10
Angela: I mean what's the point talk them down two bucks per book it's just that's a telltale they're plucking up the world's remaining rare and out of print books be before their competitors can they don't have time to negotiate.
32:20
Angela: Yeah they want that text before anybody else can do what everyone's asking wow and once all those books have been ripped apart scanned and pulped they're gone forever.
32:29
Brian: Yeah all right so let's change gears for a minute here um so regarding Anthropic's recently settled one point five billion dollar lawsuit I want to say it was one point two but was it one point two it was a lot of money okay yeah we're close over a billion dollar lawsuit with publishers and authors um an author contacted us this week saying that three of his books were stolen and illegally used by Anthropic he's expecting a nine thousand dollar payout and that was a class action suit uh that allowed victims to opt out uh do other dozens of authors did opt out with prominent groups filed um filing separate copyright infringement lawsuits in the U.S.
33:15
Angela: District Court for the Northern District of California and this was led by authors such as Dave Eggers and Angie Cruz and uh John Carry Carrial Thank you for saying all the hard names you're much better than I am at it I looked it up it is 1.5 billion not 1.2 billion i i had the five turned upside down in my head and turned it into a two so yeah okay so 1.5 billion that's the number we're gonna settle on today right right we're gonna settle on that settlement number yeah okay uh as of this past May 26 authors had joined that separate lawsuit they they'd opted out of the class action and decided to file their own lawsuit they're asking for a jury trial to avoid getting bottom dollar settlements through the class action lawsuit because typically class action lawsuits people get almost nothing.
34:08
Angela: Now the fact that anthropics victims are getting $3,000 per book is astronomical.
34:13
Angela: I've talked about this before I I was part of some class action lawsuit and I got a check for 25 cents which was less than the cost of the stamp that it was mailed with I can tell you I've never joined a class action lawsuit since then.
34:25
Angela: So again it's estimated that each author of the Anthropic settlement is going to get around $3,000 for each copy of their ebook that was illegally downloaded and each book that was illegally downloaded and used to train their system.
34:39
Angela: There's a database online showing which books were stolen by Anthropic.
34:43
Angela: We'll put a link to that under this episode right as we podcast episode 51 I checked and months ago when they first put that up I checked I didn't find any of my books darn it uh um I also looked for booklocker books our authors booklocker books didn't find any of those either but then that guy contacted us this week and there were three of his books that were taken not one of our authors if you find we'll put a link to that database underneath this episode if you find any of your books in there and if you didn't specifically opt out of that lawsuit before February 9th 2026 that was the deadline you can't opt out now.
35:18
Angela: You were automatically included in the class however you also had to file a claim so so if you find your book in there today and you never opted out and you didn't file a claim it's too late.
35:30
Angela: You would have needed to file that claim by March 30th 2026 and at Writers Weekly we announced to our readers about this lawsuit months and months ago so yeah gotta read the website it's also been all over the news I would be surprised if if any authors listening to this were not aware aware of that lawsuit months ago.
35:49
Angela: Yeah um when I did the research had any of my books been in that database I would have opted out the class action suit because that 25 cent stamp I mean that 25 cent check I got yeah I would have hired my own lawyer.
36:03
Angela: And for those who did opt out of the class action they may end up that may end up costing Anthropic far more.
36:10
Brian: Yeah.
36:11
Brian: I wonder how I wonder what their what their threshold is that they've calculated that they can afford to pay out in lawsuits and fines and whatnot to make still still make this profitable.
36:23
Angela: Well again I don't think they counted on being sued in other countries and that's right around the corner.
36:28
Brian: Okay for those who did opt out the civil penalties against Anthropic could and likely will include statutory damages and these are standard fines set by a judge from $750 to $30,000 for each work used without permission.
36:44
Brian: Another penalty could be willful damages and this is when Anthropic did so up to $15 I'm sorry $150,000 per work if the court finds the person or company broke the law on purpose.
36:58
Brian: They could also be charged court costs payment for the owner's lawyer fees and the court expenses.
37:04
Angela: It's gonna be expensive now think about that instead of a measly three thousand dollars per work the authors in the separate case could it get up to $180,000 per infringed work plus attorney's fees.
37:19
Brian: I could deal with that yeah right sorry my book wasn't on there I could use an extra $180,000 in my bank account okay so let's talk about what could happen to anthropic and other AI companies if the government pursues criminal charges.
37:33
Brian: So they could be fined up to $250,000 per offense for serious or commercial piracy and imprisonment which I don't think's gonna happen but up to five years in prison for first time felony offenders and up to 10 years for repeat offenders.
37:51
Angela: Well I definitely think the commercial piracy is is exactly what they did.
37:55
Brian: Yeah the problem is is when it comes to a company doing that the it's it's difficult to when it comes to criminal stuff digging down and in and singling out individual people to say okay you're the one that's going to bear the burden for this entire company's operation.
38:13
Brian: Right uh and and to go to prison for that it's it's it's kind of tough.
38:18
Angela: I really wish they could figure out which personnel were responsible and put some people from Anthropic and other AA companies into pokey.
38:25
Angela: Yeah because these were these were egregious acts intentionally egregious acts uh you know I'm pretty certain anthropic's gonna settle with this new group of authors as well but again they're gonna each author's gonna end up making a lot more money than 3,000.
38:38
Angela: Yeah.
38:39
Brian: Okay so what can authors and book lovers do about this situation besides keeping old and rare books that they have already that they already have on hand.
38:48
Angela: Yeah right I'm not giving up any of my old books.
38:50
Angela: You know remember we lived in Florida that awesome used bookstore?
38:53
Angela: I can't remember the name of it because don't say it out loud because AI might be listening.
38:58
Brian: But right there in St.
38:58
Brian: Pete we had that little old well not well actually it wasn't very little it was pretty big.
39:02
Angela: It's a big used bookstore used books oh it was just a gold mine we go in there and spend hours and we take the kids and everything.
39:08
Angela: Just browsing things you know if we still live there hey we're gonna be there next year next month we are going to be there next year we should go pluck up some old used books out of print books and save for the grandkids someday.
39:18
Brian: Yeah uh one thing you can do is lobby the government to update existing copyright law and bring it in line with current technology and this isn't difficult.
39:32
Angela: No it's not difficult at all.
39:33
Brian: We'll provide another link under this episode at writersweekly.com slash podcast episode 51 uh where you can find your represent representatives contact information simply by typing in your zip code um you have to contact the United States Congress specifically your local U.S.
39:52
Brian: representative and U.S.
39:53
Brian: senators as they hold the constitutional power to write and pass legislation.
39:58
Angela: Yeah it's very important for the entire literary community right now today to come together with us and everybody else lobbying against AI companies and stop them from taking advantage of a law that was created more than a century before the internet ebook and digital digital processes existed.
40:14
Brian: Yeah it's this has to be changed right now it doesn't make sense that we'd still be operating under something that just it doesn't even touch on all the all the possibilities that are available now.
40:24
Angela: It's you know it's not only is it it makes me so mad but it hurts my heart yeah to think of all these books just gone forever.
40:34
Angela: Yeah yeah they're never coming back.
40:36
Brian: Yeah and uh we'll certainly keep you updated on the new and increasing numbers of lawsuits being filed against AI companies for copyright infringement.
40:46
Brian: Yeah we'll be be posting those on keeping up keeping everybody up to date on writersweekly.com okay yeah and well I mean it's it's again it's just all you can do is you you need you need to get the the laws up to date on on this kind of thing.
41:04
Brian: Um because yeah technology is just is just um changing so so fast uh at such a rate that you know the old laws just don't apply anymore.
41:18
Angela: I honestly can't believe that old part of the copyright law still existed from 1908.
41:21
Angela: They did do a slight update in 1972 when they were updating the length of time that works could remain in copyright before they went into public domain.
41:32
Angela: Yeah.
41:32
Brian: But that part stayed yeah and I it just blows my mind you know and and I just I I I can't help but wonder what what are they what do the AI companies want to do with all of this knowledge and information.
41:48
Angela: What what is it what is it you're gonna They want to have the biggest most robust AI systems so everybody who used those instead of others I'm still trying to figure out how they're making money.
42:00
Angela: I mean I'm like at what point do you actually make money yeah right yeah yeah we I um I was on a local group last night and I posted this is this is how some people are so uh dismissive of everything that's happening in the world they're about to build an AI center here in Georgia not near us but it's a couple hundred miles away and I posted to the local group um get the word out about this so more people know about it so people can fight against it with their city council members and stuff.
42:29
Angela: And one woman wrote and said well that's not even our town and I said if everybody in Georgia doesn't come out to help the little towns that are that are having these AI data centers put in you know it'll be it could be us next.
42:41
Angela: Yeah you know we all need to come out and be vocal about this.
42:44
Brian: Well I'm not completely against AI data centers but they need to be it needs to be balanced and it and you know if if you're gonna build one there needs to be contributions to the community you need to make sure you're not using too much electricity and water and all that kind of stuff.
42:58
Brian: You know it needs to be a benefit not a drain.
43:02
Angela: Well I think they need to be out in the middle of nowhere because I would prefer they be out in the middle of nowhere.
43:07
Angela: People are reporting that um their cows stop calving yeah um uh the bees disappear there's a lot of wildlife problems with the data centers they make this constant sound that agitates animals and humans and probably the most bizarre thing is that these data centers are moving in and people's electricity and water bills are going up.
43:31
Brian: That's the kind of stuff that need that towns need to stand up and say okay if you're gonna build this here here's what's gonna happen you're gonna contribute X amount to the overall grid and the power that's generated you're gonna be paying you know for this percentage of everything.
43:46
Brian: Yeah.
43:47
Brian: Seeing as you're gonna make all this money off this data center, you need to give back in some of these communities.
43:52
Brian: Somebody else's electricity should not go up.
43:55
Angela: Yeah to get the AI data center in there and and again residents aren't typically informed of this And they say, Well, okay, we won't make you pay for electricity and water for five years because you're going to bring so much revenue into the area.
44:07
Angela: They don't inform the residents that, oh, guess who's going to be paying that?
44:10
Angela: You know, people have seen their electrical water bills triple and quadruple because and they had no idea that they were going to be responsible for paying another.
44:24
Brian: Right.
44:24
Brian: And there needs to be um you know compensation to the communities and make sure there's not going to be any kind of impact.
44:30
Brian: Exactly.
44:31
Angela: And I think it can be done.
44:33
Brian: Why why don't they have to pay to create their own electric I think they're thinking in the old model of, you know, like let's say you want an Amazon distribution center.
44:44
Brian: So hey, we're not gonna we're not gonna make you pay property taxes for the first ten years.
44:48
Brian: We're not gonna, you know, make you we're gonna give you a discount on your electric or whatever, because they know that, you know, 10,000 of their citizens can now get a job in there.
45:00
Brian: And you know, you're gonna be bringing, you know, attracting more businesses into your into your area, which can bring more revenue in for the city 'cause they know, you know, they're they're the the cities or towns or counties they know full well they're gonna make money off of this Amazon center.
45:17
Brian: Well, I think so they give them a bunch of breaks, but the the data centers are a little different animal.
45:22
Angela: I think that they're counting on like sales tax revenue for whatever the data center creates and revenue, it's taxable revenue.
45:29
Angela: From what I've researched, those data centers run on a shoestring of humans or it's all automated.
45:35
Angela: There aren't a thousand jobs coming into the town.
45:42
Brian: It it holds data.
45:44
Brian: That's what it does.
45:45
Brian: So you have to change the model to fit what you're bringing into your community.
45:51
Brian: And it doesn't work like you know a manufacturing plan.
45:54
Brian: It's different.
45:55
Brian: So you gotta figure out how how it's gonna balance all out.
45:58
Angela: But if if you've got a storefront in, you know, Trenton, Georgia, and you're selling things, you're you're charging sales tax and remitting that to the city, it also goes to the state and the county.
46:08
Angela: But if the building running your electricity for your store, creating the electricity for your store, or these uh what do you call the storage ponds for water?
46:20
Angela: Uh retention ponds.
46:21
Angela: Retention pond running your store.
46:23
Angela: I know this doesn't make any sense because it's just a store, but just bear with me.
46:26
Angela: So let's say you're on the edge of Trenton and the electric building and the retention pond are just over the the border to the next town, that location is not creating revenue.
46:41
Angela: So that county's not getting any sales tax out of that.
46:46
Brian: Well, again, unless there is some agreement that, you know, maybe some of that some of that electricity is gonna be pumped into the grid that's feeds Trenton, whatever, you know.
46:54
Brian: Could be property taxes they're collecting.
46:56
Brian: I don't know.
46:57
Brian: But there's a number of ways to do it.
46:59
Angela: But again, charging other people for a third car private companies water and electricity is just mind-boggling to me.
47:06
Angela: Yeah.
47:06
Brian: Well, I mean, we don't like the way AI is being used in a in a lot of ways, but whether we like it or not, AI is here, it's here to stay, it's not going away.
47:16
Brian: And I think the the only way that we control this from getting out of hand is more and more people need to educate themselves and be aware and be involved in uh getting their voices heard of, you know, especially in your in your immediate area, your local area, about what you know, what you will and won't accept.
47:37
Brian: Right, what's acceptable and what's not.
47:41
Angela: Clucking up uh rare and old books and destroying them.
47:44
Angela: Completely unacceptable to everybody in the publishing and reading and writing industry.
47:50
Brian: Hey, what do you think about AI AI?
47:52
Brian: Why don't you click the comment button and let us know down at the uh well, actually on the top and the bottom of this page on writersweekly.com slash podcast episode 51.
48:02
Angela: I always stutter when I say that, and you always say it perfectly.
48:04
Angela: That's because I'm a pro now on when I need to say it, I'm just gonna point at you.
48:07
Angela: I'm a professional.
48:08
Angela: You say it.
48:10
Brian: Then people not got and if you want somebody in your corner to fight against the big AI companies from chopping up your book, come over to booklocker.com and publish your book with us.
48:21
Brian: Uh we've been in business 27 years, family owned and operated company.
48:25
Brian: We've got uh A plus rating on Better Business Bureau and five-star ratings and a lot of other places.
48:32
Brian: So uh come to us, you'll you'll we'll treat you right.
48:36
Brian: And um, Angie, what do we got upcoming in the next few episodes?
48:40
Angela: Uh next week we're gonna have Clayton Jones on.
48:43
Brian: No, we're not.
48:46
Angela: Oh, that's right.
48:46
Angela: That's the week after.
48:47
Angela: Yeah, it's gonna be a good one.
48:48
Angela: Next week, we are with Troy and Stacy at Pros and Pro Pros with Pros.
48:52
Brian: Pros talking pros.
48:53
Brian: That's it.
48:54
Brian: That's right.
48:55
Angela: They had me on their podcast a few months ago, and they are a riot.
48:59
Brian: And we're gonna be talking a little bit more about scammers.
49:01
Brian: Yeah.
49:01
Brian: Uh what we wanted in in the next episode was um, because you all you all know our take on scammers and you all know our our knowledge of scammers, but we wanted to bring in two other publishing professionals that will kind of give us their point of view on what they're seeing as far as scammers.
49:22
Brian: Right.
49:22
Brian: So I think it's gonna be a really interesting, uh, really interesting episode.
49:26
Angela: And they're both hilarious.
49:27
Brian: That's right.
49:27
Brian: Yeah, you know, and then then what do you got coming up after that?
49:30
Angela: After that is uh Ten Things Authors Unknowingly Do That Break the Law.
49:35
Brian: Ooh.
49:36
Brian: Yeah.
49:36
Brian: That sounds good.
49:37
Angela: Yeah, I actually that's the title, and I've already got I'm already up to 15, so I might have to change the title.
49:42
Angela: Do some consolidation, but um, some juicy stuff in there, and it'll it that'll be a funny one.
49:47
Brian: So all right, y'all.
49:48
Brian: So keep listening.
49:49
Brian: We got good stuff for you coming up, and we'll see you again next time on the Writers Weekly Podcast.
49:57
Brian: Everybody have a great weekend.
49:59
Brian: Bye.