From managing data to performing routine calculations, AI has demonstrated its ability to free us from mundane responsibilities. However, as technology continues to advance, a thought-provoking question emerges. What happens when we train AI to replicate the intricate work of creatives, like authors? Recently, Singapore made headlines when production company Dear Straight People announced that the script for its BL series was created by ChatGPT.
As AI-generated writing gains traction, it not only shines a spotlight on the legal issues surrounding creative emulation but also prompts reflection on the future landscape for authors and other artists.
AI-Generated Works, Copyright, and Piracy
When author Jane Friedman was on Amazon, she found that there were about half a dozen books published under her name. But she did not write them. An AI algorithm did.
And she isn’t the only author who’s had this happen to, as other authors shared similar experiences.
Apparently, scammers are producing books by training AI systems to emulate an author’s style, and then get these bogus books on Amazon and Goodreads, credited to legitimate authors.
This isn’t as simple as piracy (which, for some, is considered a good thing for publicity). This is some huckster claiming to be a certain author, and profiting off the author’s name with AI-generated work. This does more than just stealing revenue – it could damage an author’s reputation with subpar work.
Everyone is asking: isn’t this illegal, and couldn’t the sites take down those AI-generated books? It’s not that simple, because this issue reveals a huge legal loophole in AI-generated content that could pose serious problems for the book publishing world.
What is fair use and what is theft?
When an author produces a book, their work is automatically copyrighted to them. This means that no one can copy it without paying royalties; doing so would be piracy. However, the issue at hand is not that AI copied Jane’s work. Rather, it was trained to emulate her writing style, leading to AI-produced content being presented as if authored by her.
So when Jane tried to get those books off Amazon, she was told:
- to prove that she held the copyrights to those books. Here’s the thing: any work produced by AI doesn’t have any copyright. And since Jane didn’t write those books, they’re not considered her works, so the copyright law doesn’t apply.
- to show that her name was trademarked, which meant that if someone tried to steal her name to write something, they’re breaking the law. This works similarly to designer brands, where you can copy a design or style, but not their brand names. Jane didn’t trademark her name.
Also, anyone can open an Amazon account and sell those books, which means that there can be countless accounts selling different versions of AI-generated work. And doing so isn’t against the law (yet). Besides, Amazon gets a piece of the revenue from every sale, so naturally it would prefer to keep the titles on their website.
Jane’s AI-written books were also listed on Goodreads (a social platform for book enthusiasts) under her profile. Interestingly, authors aren’t the ones responsible for the listings on their page – the task goes to a group of volunteer “librarians”. So, to take those books down, Jane was required to join a group and then post in a comment thread to have the illegitimate books removed.
The scary ways AI is being used
With the increasing availability of AI tools, numerous websites are now utilising AI to scrape information from the internet. Recently, the owner of a site called Prosecraft admitted that its AI had been trained on 25,000 novels scraped off the web. These novels are copyrighted works, and none of the authors gave consent. So the site was shut down after several notable authors spoke up.
Unfortunately, when it comes to written work, scammers are blanketing all genres with their AI-generated books. On Amazon, it’s not just novels or self-help works being produced by AI. There are books on “travel, cooking, programming, gardening, business, crafts, medicine, religion, and mathematics” according to a New York Times report in an article on AI-generated content proliferating Amazon. Part of the success of these scams is that sellers tend to keep prices low, with phony 5-star reviews.
AI-generated titles are also proliferating on Amazon for another reason: Kindle Unlimited.
No real humans are buying these bogus titles. Scammers are using bots to read these AI-generated titles through Kindle Unlimited to generate income. Kindle Unlimited is a subscription service where readers can borrow and read books, and authors are compensated based on the number of pages read (measured as Kindle Edition Normalized Pages, or KENP).
Bots read these fake novels through KENP, where each page read attracts a small payout. But with thousands of books and hundreds of bots, it quickly becomes millions of pages. These bots can potentially impact Kindle Unlimited earnings for real authors in a negative way by artificially inflating the number of pages read by simulating borrowed books.
Moving forward with AI
If you’re a writer who wants to know what recourse to take should this happen to you, here’s what ChatGPT suggests:
Copyright Registration: Authors can register their original works with copyright offices, providing legal evidence of their authorship.
Publicly Documenting Creation Process: Authors can document their creative process publicly (ie. on social media) providing a timeline of their work.
Monitor Online Platforms: Regularly search for instances of AI-generated content that might be falsely attributed to them on various platforms.
Engage Legal Support: In cases of infringement, authors can seek legal assistance to issue takedown notices or pursue legal action against those distributing AI-generated content without permission.
Educate Readers: Authors can educate their readers and fans about the risks of AI-generated content.
Leverage Creative Communities: Engaging with writing communities for support to raise awareness about the misuse of an author’s work.
Dialogue with Publishers and Platforms: Authors can communicate with publishers, distributors, and platforms to raise awareness of the issue.
The good news for Jane is that AI-generated titles with her name appear to have been removed from Amazon and Goodreads. It’s probably in no small part due to her visibility and reputation in the writing and publishing community. But what about when this happens to authors with smaller profiles and less resources?