Latest News

The recent controversy concerns a massive archive of 7.5 million pirated novels and 18 million research articles, called LibGen. It is alleged in an article in The Atlantic that Meta (owner of Facebook, Instagram and WhatsApp), decided that rather than pay licences for this material, they would simply download and incorporate this illegally pirated, copyright material into their Large Language Models. They basically circumvented the established methods for obtaining such permission to save money and time. Allegedly, the green light for this came from the very top, ie Mark Zuckerberg.

All 13 of my DCI Warren Jones novels are included in the dataset.

The aim is to allow AI chat bots to generate new material in the style of previously established authors, using existing characters, with no permission (or compensation) offered to the owner of such material. Furthermore, the models will learn from the writing skills of authors, honed over many years, to generate new works to rival those of hardworking, skilled writers. Potentially, users of the model will simply be able to ask the bot to ‘write me a story about subject X featuring these characters, in the style of author Y‘. They may keep it for their own pleasure, or package it and distribute it as an eBook, either for free or a small fee.

This is a threat to the whole ecosystem of writers.

Why, I hear you ask, should a mid-list author such as me (and 99% of other writers) be concerned that folks are generating new Jack Reacher novels for their own pleasure? Millions of readers will still rush out to buy the latest Reacher novel. The bank balances of superstar authors will probably only take a small hit.

But consider this. I am a typical, avid reader. Over the course of a year, I have a number of go-to authors, such as Lee Child, Michael Connelly, Karin Slaughter etc. I will always buy their latest book. But between those tent-pole releases, I read many other books. Books by other, mid-list writers such as myself. I try new novels that don’t bother the bestseller charts, written by authors whose writing income is usually only a part of their livelihood, because median author earnings (ie most of us) are less than minimum wage.

Now let’s assume somebody uses AI to generate 5 new Reacher novels, releasing them for pennies or even free. Sure, they aren’t a patch on the real thing, written by Lee or Andrew Child. But they contain all the elements of a Reacher novel, they are fun and they provide a Reacher fix until the next one is released by the Childs.

But whilst I’m getting my AI Reacher fix, what am I NOT reading? Which 5 mid-list authors will I never discover? Whose sales figures will be impacted because there are only so many books a person can read in a year? Which authors will have to slow their output, or even quit altogether, because their writing income is unsustainable and they have to spend more time doing other jobs?

That is something we should all worry about.

I am less concerned that Amazon will be flooded with AI knock-offs of my DCI Warren Jones characters, than I am that fans of police procedurals who have finished their favourite author’s output won’t give mine a go, because they can generate a brand-new DCI Banks, or Roy Grace or John Rebis novel for free, or download a new adventure of their favourite character for pennies.

Other authors are NOT my competition, they never have been. We promote each other’s work because we are fans of each other. Book bloggers and keen readers flood social media with recommendations of similar authors. If I go onto any Facebook readers group and say that I’ve read all of a particular author’s books and fancy reading something similar, the comments will be flooded with helpful suggestions. Sometimes I am flattered to be included in the same post as other writers I admire. I have gained many new readers this way. How many potential new readers will instead stick with what they know, rather than taking a punt on someone new?

And that is why AI generative novels are a threat to us all.