The Authors Guild – the Society of Authors’ counterpart in the US – has published an open letter to generative AI Leaders, calling for consent, credit and fair compensation for all historic and future uses of copyright materials in the training of AI systems. The letter has already been signed by more than 10,000 authors.
Addressing the CEOs of OpenAI, Alphabet, Meta Stability AI, IBM and Microsoft, the letter criticises the ‘inherent injustice’ of the development of these systems, such as ChatGPT and Bard. It says ‘Millions of copyrighted books, articles, essays, and poetry serve as the “food” for AI systems’, yet authors have not received any compensation for their contributions.
In their press release, they go on to say, ‘Where AI companies like to say that their machines simply “read” the texts that they are trained on, this is inaccurate anthropomorphizing. Rather, they copy the texts into the software itself and then reproduce them again and again’. The outputs of large language models (LLMs) rely heavily on authors’ ‘language, stories, style, and ideas’, without which they would be ‘banal and extremely limited’, however this is an unlawful use of authors’ works.
The SoA has already been very vocal about such unlawful training and the threats to author income posed by some AI applications. KPMG has recently estimated that 43% of the work of authors, writers and translators could be replaced by generative AI.
The letter demands that developers of AI should obtain permission for the use of copyright materials and that writers should be duly compensated for past and ongoing use of their works. It also states that writers must be fairly compensated for the use of their works in AI output, whether or not the outputs are infringing under current legislation.
The Society of Authors wholly supports the letter’s demands. We are also pushing for transparency in relation to sources, credit to authors whose work is used, the right to remove works from such systems, transparency and full disclosure whenever work is AI generated and protection from deep fakes.
Unregulated and unfettered, AI poses an existential risk to authors in the immediate and long-term, but it doesn’t have to be like this. AI developers must work together with authors to protect their livelihoods, which are the cornerstone of the publishing industry and of successful AI systems of the future.
Another problem with this letter is that it acts as if the USA is the most important country in the world and that if we beg American ‘leaders’ to treat authors better, then all will be well. This is such a narrow way of viewing what is currently happening in the world. Mark Zuckerberg and his friends are inventors, nothing more. They are not gods. And even if they reined AI in, inventors in Russia, China, Saudi Arabia, Japan, etc, etc, would still be able to invent these things. We’d still have the same problem on our hands. The only… Read more »
This letter makes me feel like I am being asked to crawl along the floor amidst the pillars of the Throne Hall of Persepolis to beg the great Shah of Persia to stop treating us little people so badly. I do not consider people like Mark Zuckerberg to be people I need to plead to. And there are, I think, better ways of stopping AI hurting us all. I see AI like the image of people tied to the machines in the Matrix. There is no point trying to get people at the top to stop AI; we have to… Read more »
The red pill, Neo. No. The red. Not the… Oh, Neo. Such a disappointment.
I don’t understand your reply, Morpheus. You were so much better at explaining yourself back in the 1990s when you did so at length to Neo. You wouldn’t be the only one who’s lost the ability to write at length, of course, if that is what the problem is here. Twitter made everyone, for a long while, unable to read more than 280 characters in one go. How this damaged the art of deep and reflective reading – and this was well in advance of all of this fuss about AI. Fortunately this war between Twitter and Threads appears to… Read more »
This AI is very dangerous and potentially damaging to a wide variety of both fiction and non-fiction authors as well people active in all media.
When so many of the fathers of AI appear to be either engaged in a race to be first onto the marketplace with newer and bigger and faster versions of this barely understood technology or issuing warnings of the cataclysm to come, it’s hard to remain sanguine about the future harms or benefits of this man-made intelligence. I’m less concerned about having my livelihood rapidly eroded and destroyed as I am about our civilisation being ended by an intelligence that cannot be aligned with human ethics and aspirations.