If you follow historians or authors in general on social media, you will be aware of some of the fury that has been posted in the last week since The Atlantic’s Alex Reisner shared his investigation into the ABHORRENCE that is LibGen, a pirated book database originating in Russia that Meta (Facebook, Instagram, Threads, Whatsapp) are using to train their flagship AI model, Llama 3.
The gist is this – Meta’s AI has been trained by its engineers to access LibGen’s catalogue of pirated work to build up its knowledge, claiming ‘fair use’. LibGen is known as a shadow library, an online repository that holds up to 7.5m books and rising. Effectively, it provides access to copyrighted works, and it is this that has been used by Meta for its generative AI.
Meta’s AI system is, by Mark Zuckerberg’s own claims, embedded into the Meta system and used by hundreds of millions of people worldwide, including one would expect you and I. Many other AI companies presumably can and have also tapped into LibGen to feed into their systems. This navigates around the more costly, slower legal process of licensing books through the traditional route. The Atlantic calls this piracy and that it is illegal.
The use of AI systems, as the Writers Guild points out, has been ‘gathering at pace in recent years’, and much of the data used to train these systems are being mined from the work of creatives, unknowingly, without consent, and illegally.
Authors produce books through a deep-seated labour of love for writing, an art built upon expertise and education, works that take many years to produce and which, for the overwhelming majority, do not necessarily make much money. It was recently revealed in one academic survey that the average author is only earning around £7,000 a year from their work, and that just 19% of writers earn their income solely from writing. There is very little money in this game.
So, to find out your work is being STOLEN without recompense is naturally INFURIATING and DEEPLY DEMORALISING. Original work should be read by as wide an audience as possible, and most authors will be keen to support diversity and inclusion initiatives. But the people who are producing this work cannot expect to do so for free, to be so callously exploited. Would you work for free?
WHAT CAN YOU AS AN AUTHOR DO?
Though it feels very much like the horse has already bolted, there are attempts by the industry to fight back against this theft. First and foremost, you can search The Atlantic’s LibGen database to see if your books have been stolen.
The Authors Guild have produced an article saying they are collaborating with publishers and the US Federal Government to try and combat these piracy sites, though admit they struggle to do so as many are based in Russia or outside US jurisdiction. They say legal action is underway in California and will update with the outcome this summer. The Authors Guild also provide details in their post about sending a formal notice to META et al with a template they have created, and how to add a ‘NO AI TRAINING’ copyright to your work (this may necessitate publisher input I’d expect for trad published)
In the UK, The Society of Authors have also been active. They are clear that what Meta have done is, under UK fair dealing law, unlawful and are calling for right-holders to be compensated. The SoA is, therefore, campaigning for increased protections for authors and to end AI companies stealing copyrighted works without payment.
The SoA are raising the matter with Meta and the UK Government, but also implore you to tell your publisher and agent and to pressure your local MP by writing a letter to safeguard creative rights. The Creative Rights in AI Coalition have an email form you can use, that requires simply putting in your Postcode to find your MP.
The Writers Guild have also issued a strongly-worded statement with advice for authors, which follows much of what has already been covered above. Helpfully, they have put together a comprehensive page covering the issues with AI and how it poses significant threats to writers’ work and earnings.
PROTECT OUR CREATIVE RIGHTS
The bottom line here is that Meta and other AI companies are stealing authors’ creative works, and doing so without permission, citation, credit, or payment. Something must change before this becomes the norm. Why should people continue working in creative industries if tech billionaires alone are the beneficiaries.
How do writers pay their bills, their rent, their food if this is where the industries are going? Will our government intervene and protect us through harsher regulation? I won’t hold my breath. Big Tech has western democracy by the throat, these days.
AI can do great things for humankind, but in the wrong hands, it is a poison. And right now, it is infecting the publishing industry and stealing the livelihoods of authors.
Well said Nathen. As mentioned earlier please let us know how we can help/support
A very important article, Nathen. It’s appalling how authors are being treated.