Lise Hand: AI companies destroying books to feed the machines? What fresh hell...

Lise Hand Journalist-at-large

IT’S NOT OFTEN these days that a news story will cause my jaw to hurtle floor-wards as an outward expression of inner stupefaction. Not because the news is any less terrible than it has been at any point since, oh the last century, what with plagues, wars, a burning planet, a demented cult dismantling American democracy and so on. It’s more a case of being in that state of insensible desensitisation experienced by the frog chilling in what it thought was a bubbling whirlpool, while idly observing that the water is feckin’ roasting today.

On the face of it, it was a nobody-died sort of article. The headline was even punctuated with a question-mark, which generally doesn’t signal hair-on-fire panic. Yet suddenly I was in the grip of the sort of creepy feeling usually portrayed in apocalyptic films by a grizzled farmer staring at an empty sky and wondering aloud where all the birds have gone.

“Why is Anthropic destroying books?” asked The Guardian last week.

It wasn’t the first piece written on this relatively recent phenomenon, but it made for stark reading. It highlighted court documents relating to a North California district court copyright case, Bartz v Anthropic PBC, specifically a memo on ‘Project Panama’, one of the AI company’s efforts to improve its large language model Claude. An internal memo explained what the project entails: “Project Panama is our effort to destructively scan all the books in the world.” The memo advises discretion: “Why use a codename? … [B]ecause we don’t want it to be known that we are working on this.”

It’s a complicated yarn, outlining the options considered by the company to get ahold of all the books in the world. Many options were legally fraught and time-consuming, and the march of tech does not wait for any man, woman or child. Therefore – money not being an issue of course – Anthropic embarked on a buying spree, snapping up print copies of every book available, in order to “destructively” scan them to electronic form. As the court’s decision relates, the company’s vendors “stripped the books from their bindings, cut their pages to size and scanned the books into digital form, discarding the paper originals.”

Sucking up knowledge

For the past couple of years, stories have been popping up in various countries of bemused booksellers receiving massive bulk orders for stock, often including odd or outdated titles. In April, Galway-based Kenny’s bookshop told of a “bananas” order they had received for 5,000 titles, which included a range of obscure and eclectic books.

This ‘repurposing’ of books from brain food to LLM-fodder was ruled legal under US copyright laws – hence the mass shipping of container-loads of bulk-bought books from Europe to America. But the fact that the courts ruled what it called this “transformative” process as legally kosher doesn’t detract from the hubristic horror of it all.

Advertisement

Maybe this personal sense of existential dread is simply the massive overreaction of a lifelong bookworm. As a kiddo, I read by torchlight under the blankets after lights-out. Once I had read the entire works of Enid Blyton – which sparked the formation of various neighbourhood gangs and also a mercifully unfulfilled longing to be packed off to a boarding-school – I hit the local library. There was no YA – you read kids’ books, and then you read adults’ books. I’d hoover up anything – science fiction, crime, historical fiction, the classics, stuff involving pirates or badly-behaved heroines.

In convent school, a copy of Coleen McCullough’s risque romance ‘The Thorn Birds’ whipped surreptitiously around the classes like a coronavirus, in the days before we knew that a priest having sex with a young woman wasn’t just outlandish fiction.

Studying English literature in college was a no-brainer. I fell in love with Flann O’Brien and took an unshakeable set against DH Lawrence. I discovered Andre Gide and Jean Genet and briefly succumbed to pretentiousness. For a while after education, books were replaced by the twin distractions of music and boys, but inevitably fought their way back into my affections.

How could they not? I was forever being astounded by novels. I was astounded by the so-called hard-boiled writing of Raymond Chandler, who could transform one gruff sentence into a whole tender short story. I was ROCKED by the opening line of Hunter S. Thompson’s ‘Fear and Loathing in Las Vegas. (“We were somewhere around Barstow on the edge of the desert when the drugs began to take hold”. Like, wtf?).

I slept with the light on for weeks after reading ‘Red Dragon’ by Thomas Harris. I tore through ‘A Handmaid’s Tale’ by Margaret Atwood and from the innocence of the 1980s, thanked all gods that it was pure sci-fi fantasy. Gripped by Maya Angelou, Virginia Woolf and Simone de Beauvoir, I became an enemy of the patriarchal status quo. On a whim, I picked up Hillary Mantel’s new book ‘Wolf Hall’ in a New York airport and read it nonstop on the flight home and all through the next day and following night, dazzled with the apparent ease with which she scattered language like handfuls of rubies and sapphires onto the pages and conjured up the past like a witch. I’ve been haunted by books, but also often irritated by them – it took years of spine-stiffening before I was able to abandon one half-read without feeling absurdly apologetic to the damned thing.

Bookworms and bibliophiles don’t dwell in towers of academe; we can read anytime, anywhere, and we know a thing or two about the human condition. Give me five minutes alone with any potential partner’s bookcase, and I’ll have a fair sketch of their heart. No bookcase? Swipe left and move swiftly along. Petrichor is all fine and dandy, but biblichor is our jam.

And a book is a book, no matter how it’s experienced. eReaders are a gift for travellers, BookTok is a brilliant resource, book clubs are marvellous proof that reading can be both a solitary pursuit and a bonding activity, and the current audiobook boom is proof that we can read with our ears – though there are indications that the eyes still hold the advantage. Some neuroscientists argue that when it comes to understanding a story, the brain draws on many of the same language networks whether the words are conveyed via the eyes or the ears; others insist that the slow, reflective process of print book perusal allows us to analyse and reflect on a text in a manner that digital media struggles to replicate.

Related Reads

Lise Hand: Gotta hand it to Donald Trump, he’s bringing protest back

Ireland should embrace being the land of the piseog

Lise Hand: EnBritification - has Britain just run out of luck?

One question: why?

So why is at least one AI company destroying books? History has repeatedly shown that nothing good has ever happened when regimes obliterate reading material, as did Glavit in the Soviet Union, China’s Red Guard during the Cultural Revolution and book burnings in Nazi Germany. In fact, more than one classic novel has issued dire warnings from the past about the dangers of such actions – Ray Bradbury’s ‘Fahrenheit 451’ in which Firemen burn books to suppress independent thought and maintain social conformity, and (inevitably) Orwell’s ‘1984’ which features state-enforced rewriting of history and the destruction of unorthodox materials because “who controls the past controls the future. Who controls the present controls the past.”

But, but, goes the counter-argument – these eviscerated books are for LLM training purposes only, not to rid the world of printed works, whether fiction or non-fiction. And when the machines have all the knowledge of humanity, everyone will benefit from access to such a deep well of collective wisdom.

Nope, nope, not buying it. There’s something viscerally repellent about the image of a vast warehouse where countless books are being methodically dissected to be fed into the voracious maw of a learning machine. Books may be non-sentient, non-living objects, but the ideas, ambitions, sorrows, hopes, humour, lyricism, loves and fears of their authors breathe through their pages.

What happens if, now knowing everything, the next logical step is to eradicate perceived imperfections such as texts not in tune with modern strictures? Anybody for “It is a truth universally acknowledged that a single person in possession of financial security must be in want of a life partner”? Anyone?

God almighty. Off with me to a bookshop to get a hit of biblichor. While I still can.

Lise Hand is a journalist and writer. This column runs every Wednesday.