Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

7am

7am artwork

AI is eating the world’s books

Sep 27, 202617 minAudio
  • Apple Podcasts
ShareXLinkedInFacebookWhatsAppEmail

A rare old book can now be worth more dead than alive. Books are being bought, cut apart and scanned to train artificial intelligence. Nobody knows how many books have been destroyed this way, but second-hand booksellers around the world have reported anonymous orders ranging from thousands to as many as a million books at a time. And internal documents from AI giant Anthropic revealed plans to “destructively scan all the books in the world”. Today, author Sarah Wilson on the destruction of knowledge to build artificial intelligence. If you enjoy 7am, the best way you can support us is by making a contribution at 7ampodcast.com.au/support. Socials: Stay in touch with us on Instagram Guest: Author Sarah Wilson Photo: AAP Image/Lukas Coch

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

Speaker 1That's the sound of a book being scanned to train AI. This one remains intact. Others face a more brutal fate.

Speaker 2Cut the spine off it, rip the pages out, scan them, throw it away.

Speaker 1Nobody knows how many rare old books have now been bought and destroyed to feed the AI beast. But second-hand booksellers around the world have reported anonymous orders ranging from a thousand to a million books at a time. And internal documents from AI giant Anthropic have revealed plans to, quote, destructively scan all the books in the world. I'm Ruby Jones, and you're listening to 7am. Today, author Sarah Wilson on

Speaker 1the destruction of knowledge to make artificial intelligence. It's Monday, September 28th. Sarah, I thought we could begin with this bookseller who became suspicious about who was actually putting in orders for the rare books they owned and ultimately managed to find out quite a lot about what was going on.

Speaker 3Tell me about what happened.

Speaker 4Yeah, so there's this kind of network of rare booksellers and they'd been getting these large orders for rare and unusual books, but specifically ones that were older or were published pre-2022.

Speaker 3One particular bookseller.

Speaker 4Teamed up with a media organisation, an independent media organisation called 404 Media. I had set up the AirTag, so it was linked to my phone and I could see where it was going. And they put an AirTag in one of the books and tracked this book all across the US.

Speaker 3But it left California from an airport. Next up, it appeared in an airport in Wisconsin.

Speaker 4It was in.

Speaker 3A warehouse near the airport.

Speaker 4Until it finally arrived at an Amazon warehouse in Las Vegas or the Nevada desert. And this particular warehouse has a logo. It's a picture of a Tyrannosaurus rex attacking a book. And what they were able to find is that Amazon workers had been sharing information about what was going on in this particular warehouse. They'd been charged with removing the spine of each one of these books... so that the pages could be easily photocopied, scanned and then

Speaker 4digitally uploaded. And the allegation is that these pages were being used to feed the LLMs, the large language models for Amazon.

Speaker 1And we've since learned that this isn't isolated to one bookseller or a few booksellers or even one tech company. It turns out that Anthropic has had this secret plan as well to buy and then destroy large amounts of books. So tell me about that project that they were calling Project Panama.

Speaker 4Yeah, once again, it's secretive. It was only discovered through court filings for another court case, a related court case that Anthropic was going through for piracy. And so there were a bunch of authors who did a class action seeking compensation for piracy. And in the process of that court case, it was revealed that they were also engaged in this, what's called destructive scanning.

Speaker 2It's the largest copyright settlement in American history. The AI company Anthropic forced to pay out a whopping $ 1. 5 billion after a judge said it illegally pirated more than 7 million digitized books to train its AI model.

Speaker 4And the court documents that were revealed actually found that the aim was to destructively scan all books in the world for this particular project. But once again, Ruby, we only found out about it by accident via another court case with a bunch of authors, which, by the way, I was implicated in. I am one of the authors in this class action against anthropics. So this now seems to be a widespread way to be able to scan books without actually having to pay authors copyright. It's far

Speaker 4cheaper to buy a copy of a book, get it shipped to some big warehouse somewhere and scan and destroy them.

Speaker 3And so why destroy them though?

Speaker 1So the books are being used to train AI models, but why then go ahead and get rid of them afterwards?

Speaker 4Well, it's simply easier for them. That's the real answer. So there is some speculation that in this court case with Anthropic, that the fact that they destroyed them and then replaced them with a digital copy, lended things to a fair use defense. But I should just point out that the court in the US did declare it fair use to buy a book, scan it for training purposes, and then to destroy it. The destruction is because they've already destroyed it. They've taken the spines off. And what

Speaker 4can you do with a whole heap of leaves of a book?

Speaker 1And you mentioned that all of these books that were being bought and destroyed in this way, they were written before 2022.

Speaker 3So tell me about why that is.

Speaker 4Yeah, it's significant because that was when books and materials started to have AI content in them. And so for AI to keep on learning, it needs to avoid recursive learning. This is the idea where it's actually sort of like the idea of a photocopy of a photocopy of a photocopy, you know, 10 photocopies in. the picture is deranged and unusable. That's what happens if AI feeds on its own content. So to guarantee that it's not getting its own recursive content, it has to get material from before AI was actually

Speaker 4introduced into the world, which is 2022.

Speaker 2Okay.

Speaker 1And so what happens then when these companies run out of books.

Speaker 3To use in this way?

Speaker 4Yeah. Well, this is the interesting question, right? So They've essentially scraped the internet of most of the content that they're able to use. They're now working their way through books. They have actually committed at this stage, I should say, to not using books more than a hundred years old. They're actually not that much interested in it. They actually want modern books. And it's not even something to be flattered about that my books were used because they're actually really kind of trying to get hold of weird like pamphlets, like the sewerage techniques of greater Chicago region in the 1980s. Like They're wanting really obscure stuff because they're wanting to

Speaker 4extend the learning capacity of their technology. But when they run out, that is a very good question. And it looks like they're going to run out in about 2028. That's what the modelling is suggesting. After that, they can actually have to turn to real life authors and writers to create more content, presumably. So it's a very real risk because when they do run out, their AI actually kind of starts to grind to a halt. And I think it's quite ironic that they're going to need writers. Now, this has also happened with philosophers. Philosophers are now earning

Speaker 4sort of six-figure incomes working for these tech companies because they're needed to navigate the moral maze that the technology has created for humankind. And it's quite interesting, isn't it, that it's actually these thinkers, writers, artists, creatives... The people are the very thing that AI can't create and is going to have to rely on going forward.

Speaker 1Still to come, why Sarah accepted money from Microsoft to use her books for AI training.

Speaker 3So you mentioned... that you were part of this case, Sarah.

Speaker 1So before Anthropic was buying up these books, it was using these pirated.

Speaker 3Online libraries to train its models.

Speaker 1Your books were part of those libraries, and then you were part of this group of authors who ended up settling with Anthropic for, it was $ 1.

Speaker 35 billion.

Speaker 1I mean, when you first found out that Anthropic was using your books in that way, how was that realization for you?

Speaker 4I wasn't shocked, to be honest, because I'd been witnessing this kind of thing happening over the years. My content had been used illegally without my authorization for all kinds of things. There's publishers in Iran, for instance, that take my books, scan them, translate them and produce them as physical books and just put them out to market and make money from them. There's nothing I can do because they're not part of international copyright laws. I know that my work has been all across various LLMs. for many years because people can go in there and say, please

Speaker 4write this in the style of Sarah Wilson, and it spits out literally the kind of phraseology that I've been using for years.

Speaker 3So it's in there.

Speaker 4Authors can also go into some of these pirate libraries and do a search and can find their work in there.

Speaker 3So it wasn't a surprise.

Speaker 4There's also about 75 cases that are actually out there at the moment, various authors suing other companies. So this is just the first one to come through with a settlement. And it's also the largest in history. Okay.

Speaker 1And after that class action, you were invited to opt into a licensing deal with Microsoft, which would allow them to use your books.

Speaker 3To train AI.

Speaker 1There were some authors who refused to do that on principle.

Speaker 3You decided to do it, to take the fee, though. Tell me about that decision.

Speaker 4Yeah, I did think about it for quite some time and the bulk of authors did take the fee. And I share their thoughts, but this issue is massive. There are multiple LLMs out there who are taking our content anyway. So we should take whatever money is being put forward. People should be paid. We get paid licensing fees if somebody wants to use our work for a PowerPoint presentation or in schools for lectures and this kind of thing. Now, of course, the devastating issue is that the books aren't

Speaker 4just being used by university students in a lecture. They're being used to train students a system that could replace writers, artists, creators. That is the bigger issue here. Now, is it going to be fixed by authors refusing to take the money that they can? You know, there are principled decisions that could be made here.

Speaker 3Ultimately, I decided to.

Speaker 4Take the money because I sort of saw it and went, okay, I'm taking money while I can, and I'm going to put it into funding my ability to go and fight the bastards.

Speaker 3And so how do you do that then? Once AI has gone.

Speaker 1Through all of these books, used all this knowledge, its language skills have improved as a result, how do you fight back to make sure that writers aren't made obsolete by AI?

Speaker 4Yeah, there's a couple of things. I mean, I'm talking to you right now and exposing this. We're having a conversation about it so more and more people know about it. Remember, we became aware of this because some rare booksellers started talking to each other and decided to expose it, right? This topic is now being talked about around the world. the European Commission has had discussions about copyright laws. Here in Australia, the Australian Society of Authors have also got together and are lobbying the government and have done so successfully. They've managed to ensure that there are copyright protections in place,

Speaker 4that they're not destroyed. And they're also pushing for ways to get compensation for authors going forward, but also for past compensation. So I think that is part of what needs to be done right now. We should keep in mind that, you know, These tech leaders love to talk in very, very dramatic language. We've seen a lot of that happen in the last week or so with the whistleblowers coming out. You know, so when they say that they're destructively scanning all books in the world, they're in fact destructively scanning one copy of each book. It's not

Speaker 4like they're taking all copies of my books and putting through an incinerator, but they are actually taking the content, the creative content. The other thing I would say is this is actually a sector of the broader debate that can be legislated.

Speaker 2We can actually control this.

Speaker 4We've seen it with the music industry, the Napster case over 20 years ago. That actually revolutionised the music industry and provided copyright provisions for artists. We can do the same with books. It's not difficult to do. It can be legislated against. So the more that people know about this, the more pressure is put onto governments around the world I do believe something will come out of this exposure.

Speaker 1And as you've noted, AI companies are now hiring philosophers to help them navigate this problem that they've created.

Speaker 3You've raised the prospect.

Speaker 1Of writers being hired when all of the books run out. It seems back to front though, doesn't it? To create the problem by pursuing AI and then try and get human thinkers back in to solve it. What kind of future do you think we are hurtling towards here?

Speaker 4We are hurtling to a very uncertain future. And I think what's been exposed in the last couple of weeks, what a lot of people are becoming aware of is how little the tech bros themselves know about what that future looks like. You know, they're very much stabbing in the dark. They're asking to be regulated for pauses to be put on the development of their technology. And the best thing that we can all do is be very, very aware of what's happening and And to use whatever democratic processes still remain in place to fight for, I guess,

Speaker 4retaining things that are so, so, so important to sort of human life. And I guess the worst outcome for us going forward is something that's totally Orwellian. George Orwell's 1984, The Mystery of Truth, wound up controlling the one document of recorded history and was able to change it at will to suit its party lines. Now, I don't think we're close to that yet. There's a lot of books out there in the world. And so far, as I say, the tech companies are actually committing not to buying up

Speaker 4books that are over 100 years old. We've got to ensure that stays in place. So yes, there's a dystopian future ahead of us if we don't act now I mean how many times have we heard that but I think actually this is actually viscerally disturbing anybody who's seen the imagery of these books with their spines being cut off it's actually really upsetting because books are so precious to us people preserve books we create massive monuments to books for preserving our history, 5,000 years of mindful recording

Speaker 4of the human moral endeavour. And so I do think that this particular corner of the AI travesty, we probably will be able to ensure that it doesn't go too far.

Speaker 1Yeah, books are text, but they're also symbols in a lot of different ways.

Speaker 3Absolutely. Sarah, thank you so much for speaking with me. My pleasure, Ruby.

Speaker 1Thank you. Also in the news, the federal government says it's working closely with OpenAI after one of its AI agents gained unauthorised access to Australian government websites. OpenAI boss Sam Altman and Anthropic CEO Dario Amodei have now been

Speaker 1called to appear before a Senate inquiry following similar incidents involving government sites in Australia and the United States. Prime Minister Anthony Albanese says OpenAI has questions to answer after it took almost three months to notify the government about the Medicare breach. And Iran says diplomacy is the only way forward with the United States after President Donald Trump rejected its latest proposal to reopen the Strait of Hormuz. Tehran had offered to reopen the shipping route within seven days if the US eased military pressure and lifted its

Speaker 1blockade on Iranian ports. Iran says it's still waiting for a formal response through mediators, while Trump has called for Iran's continued economic isolation.

Speaker 3I'm Ruby Jones. This is 7am. Thanks for listening.

Open the transcript file

Hosts and Guests

Guests

  • Sarah Wilson Guest Inferred

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. How Albanese killed the climate wars

    Nov 21, 202615 minAudio

    At COP31 in Türkiye, Australia has been helping lead global negotiations over the future of fossil fuels and the fight against climate change. But while the Albanese government pushes for stronger international action,...

    • Apple Podcasts

    Credit the people on this episode

  2. Ed Husic on being ‘left in the dark’ by his own government

    Nov 20, 202617 minAudio

    Anthony Albanese runs a notoriously tight ship. Across his two terms as prime minister, there’s been very little leaking, very little sniping, and very few of his Labor Party colleagues willing to speak out against him....

    • Apple Podcasts

    Credit the people on this episode

  3. Puberty blockers greenlit: So why are state bans still in place?

    Nov 19, 202615 minAudio

    This week, Australia’s leading health research body released its advice on puberty blockers, saying they should be available for some young people under 18. The Queensland and the Northern Territory governments, which...

    • Apple Podcasts

    Credit the people on this episode

  4. Will Trump, Xi and Putin try to carve up the world?

    Nov 18, 202616 minAudio

    In 1945, in the Crimean city of Yalta, three global superpowers - America, Great Britain and the Soviet Union - met to carve up the world between them, each claiming their own sphere of influence in the final year of...

    • Apple Podcasts

    Credit the people on this episode

  5. AI job cuts at Bunnings, Qantas, NAB: ‘Mass unemployment is the business plan’

    Nov 17, 202616 minAudio

    Last week it was Qantas, last month it was NAB: two giants of Australian business cutting jobs due to the growth of AI. The explanations were familiar. Qantas says it will allow the airline to unlock innovation faster...

    • Apple Podcasts

    Credit the people on this episode