On the same day my new book came out, I discovered that my first one had been stolen.
As part of its investigation into pirated material, The Atlantic has published a database that lets you search for your own work to see if it was part of the data Meta has been taking - without your consent - to train and develop AI its system, Llama 3. I was surprised at first to discover that I and my little-known poetry was likely part of this miserable, shitty act, but I shouldn’t have been. Millions of books and papers have been pulled from LibGen, an online library that contains 7.5 million books and over 80 million research papers. Almost every recently published writer will probably have had their work circulated on LibGen then - my poems weren’t special or singled out for their brilliance (sadly), simply the by-catch of the gigantic digital trawler scraping the virtual ocean.
It is not yet clear whether scraping from copyright works without permission is unlawful under the US fair use exception to copyright, but if that scraping is for commercial purposes (which what Meta is doing surely is) it cannot be fair use.
Society of Authors
On principle, I don’t actually mind places like LibGen. I don’t even mind that my poems are there - and probably soon, my stories and novels too. The most fundamental reason why most people write is to be read, to be seen, to be recognised. Getting paid for that is a critical part of the process, but for every person that cannot or will not pay, there are ones that can and do. That feels like enough for me right now. What I do have a problem with is people - specifically, huge corporations whose annual revenue is bigger than the GDPs of entire countries - using that freely available work to profit from it. Meta’s annual revenue for 2024 was USD$164 billion, up from 2023 by 22%. The main reason behind the increase? AI-driven ad targeting. Meta has used the work of millions of writers to sell things back to us.
Work taken without consent, payment or credit is, by definition, theft. Meta did all of this knowingly. The Atlantic’s investigation shows that senior figures at the company considered licensing books and research papers the legal way, but opted for stolen work instead becuase it was cheaper. And faster.
This whole thing is a legal grey area. As usual, our laws are far too slow to catch up with technology, and far less likely to skew in favour of artist, writers, musicians over corporations. According to the Society of Authors, “It is not yet clear whether scraping from copyright works without permission is unlawful under the US fair use exception to copyright, but if that scraping is for commercial purposes (which what Meta is doing surely is) it cannot be fair use. Under the UK fair dealing exception to copyright, there is no question that scraping is unlawful without permission.” There are lawsuits being brought against various AI systems by people and organisations across the creative industries, but even if they win and there are sweeping changes made to copyright law and AI, that cannot undo the damage already done. Nor can it necessarily prevent it from happening in the future.
I don’t know how I’m supposed to feel yet, other than a resigned sort of frustration. Even the Society of Authors admits there’s not much to be done. I’ve been earning money as a writer since I was 23. A small amount, to be clear, but an amount all the same. I’ve also been putting my writing out for free since then too, poems, short stories, nonfiction - theoretically available for anyone to take and do what they want with it, even if it’s mine. It’s one thing for human beings to do this - imitation is the greatest form of flattery, after all - but a robot doing it feels different. It undermines a sacred tenet of creativity: the idea that only you could have made the thing you made. That your spark, your curiosity your human-ness, your yearning and your specific experiences came together to make something specific and unique to you. I leave a piece of myself in all my fiction writing, it is my soul at this most vulernable. The idea of a robot attempting to replicate that - and using my own should to do so - feels like a violation, even if it hasn’t actually been done. The mere possibility is a threat, because it implies my eventual redundancy.
“Right now, we need writers who know the difference between production of a market commodity and the practice of an art… We who live by writing and publishing want and should demand our fair share of the proceeds; but the name of our beautiful reward isn’t profit. Its name is freedom.”
Ursula Le Guin
Le Guin was talking about the unfortunate intersection of capitalism and publishing, but this applies to the use of AI in creative writing - in all creative output - too. The sole purpose is to make more money, to undercut writers and devalue their work, to replicate it and manipulate it endlessly through a machine which doesn’t need sick days, or lengthly deadlines, or ownership, or validation, or empathy, or tenderness.
I like to think that AI will not, in the long run, triumph over the poets, the painters, the writers, the singers, the photographers, the editors. I believe that no matter how good a machine can become at creating, it is the messiness, the contradictions, the imperfections and the unpredictability of humans that makes art feel real, makes the stakes feel genuine. There’s a reason why ten million tourists a year visit the Louvre in the hopes of glimpsing the Mona Lisa, when they could arguably see it much clearer, for much longer, online. Clout aside, it’s because when we look at a human-made creation, we can sense the humanity within it. Call it whatever you want - spirit, soul, gravitas, feeling - the humanity within us responds to the humanity within art made by people. My hope remains as long as this basic truth does.
Imagery from Vika Glitter and MLK BNL on Pexels
A quick favour. I love writing these posts, and I intend to do them for free for as long as I can. If you enjoyed reading this, forward it to a friend (or three) who you think might like it too. It helps massively, because validation from strangers is truly the only thing that makes the horrors bearable for me.
Ai is much bigger than meta and copyright laws still apply