Artificial Intelligence can’t truly replace human-written fiction, figures Alan Bailey, a Pittsburgh-based writer and co-host of the podcast “If This Goes On (Don’t Panic).”
The reason is that fiction, Bailey said, requires intention. Authors must choose their words carefully and AI can only mimic others’ words and never truly create.
While Bailey can see the benefit of using the technology to tackle monotonous tasks, the lack of respect the industry shows to authors gives him — and others — pause. Over the past year, authors from Nora Roberts to Michael Chabon have voiced their discomfort with the idea that their novels could be used for profit by AI companies without appropriately crediting or compensating them. In 2023, a handful of authors filed lawsuits complaining that AI companies have been using their copyrighted materials without crediting authors or obtaining their consent.
While the parties involved fight it out in court, PublicSource asked Pittsburgh authors, publishers and a few experts what they make of AI and how it’s impacting their work. Authors agreed that the companies’ actions amounted to fraud, but lawyers say walking back the use of their materials might be easier said than done.
Authors sue, with limited success
In August 2023 the online book community bristled in response to Prosecraft, a platform dedicated to the linguistic analysis of roughly 25,000 works of fiction through AI. News that Prosecraft’s founder Benji Smith used authors’ work, without their permission, to train his platform prompted demands for removal and complaints to the Author’s Guild. Although the platform was ultimately halted with an apology from its founder, authors still fear that AI companies will profit off of their work without appropriately crediting or compensating them.
Authors filed lawsuits against AI companies such as OpenAI on the grounds of copyright infringement, arguing that there is an “inherent injustice” in AI companies using their copyrighted materials without credit or consent. The Authors Guild, which is backing one of the lawsuits, pointed out that for every author on the New York Times bestseller list, many more authors struggle to make ends meet. Using their content without consent or proper compensation, they said, is unfair.
In February, a federal judge in California dismissed most of the claims in a lawsuit filed by prominent writers who accused Meta and OpenAI of copyright infringement.
Flooded with AI sci-fi
Bailey uses AI transcription and hopes to see the technology used to ease people’s burdens, without dehumanizing them. What gives him pause is the lack of codified ethics among AI companies and the lack of respect he feels they show to authors and their work. In addition to concern over his written work, he worries that with his voice widely available online, someday it could be used by an AI company without his consent.
“If human beings are shoved to the side, if art is so fully commodified, that it’s just something that with the touch of a button you can create it through a computer program, I think that’s a loss in many, many ways to society,”
Joshua Bellin
“If AI is advanced enough, use it for the drudgery. Let it correct your spelling errors,” urged Bailey. “I’d like to see guarantees that writers, and other creatives, are fairly compensated when their work is used. Perhaps a writer should even be compensated every time an AI references their work when it is used to ‘create’ something new.”
Would-be authors send in AI-produced science fiction stories, according to Air and Nothingness Press Publisher Todd Sanders.
“These people who think that this is going to be a money-making scheme or something, or they’re too lazy to write, [or] ‘I think this is some shortcut,’ — it just doesn’t work,” said Sanders, of Pittsburgh. “And it’s getting worse and worse for us.”
For authors whose work has been digested by AI without their permission, it amounts to theft, Sanders said. Adding insult to injury, he said, is the reality that even when authors express their discomfort with their stories being used by AI, there’s no guarantee their work will be removed from a platform. Sanders, who has built his career around working with authors and artists, feels these companies’ use of authors’ work is disrespectful to the creative process.
The sentiment is shared by authors like Joshua Bellin, of Pittsburgh, who hopes lawsuit rulings empower authors to have more authority over their work. In Bellin’s ideal scenario, if an AI company wanted to use an author’s work to train its platform they’d be required to ask permission, credit the author and compensate them. It’s only fair, Bellin said, if authors’ work is used to develop profitable platforms.
Creative people believe that they have something valuable to offer that world, Bellin said. While computer programs might be able to replicate the written word, displacing the authors whose work trains such profitable platforms does a real disservice to authors.
Read more
An MS diagnosis gave a sense of relief, and now I’m figuring out how to live with it
“If human beings are shoved to the side, if art is so fully commodified, that it’s just something that with the touch of a button you can create it through a computer program, I think that’s a loss in many, many ways to society,” Bellin said.
Opting out of AI may not work
On the East Coast, the Authors Guild-backed lawsuit, which includes the likes of George R.R. Martin, Jodi Picoult and John Grisham, remains active. In January 2024, nonfiction authors such as Nicholas Basbanes and Nicholas Gage sued OpenAI and Microsoft on the grounds that the companies had used their content without consent. Most recently in August 2024, three California-based writers filed a class action lawsuit in federal court alleging that Anthropic, an artificial intelligence company, had used their work to build its platform.
There are arguments for both the authors and the companies, according to Mehtab Khan, a fellow at the Berkman Klein Center for Internet & Society at Harvard University. Copyright laws give authors certain rights, but an argument could be made for fair use, Khan said. The fair use doctrine holds that using copyrighted material for limited and “transformative” purposes can be acceptable.
Khan noted that authors can express their discontent or they can try to opt-out of AI platforms, but removing their work from those platforms might prove a challenge.
“Once a dataset is formed, and trained, and fine-tuned, it’s really hard to extract individual pieces of data that belong to a certain author,” Khan said. “So opting out has very limited efficiency.”
To mitigate authors’ indignation, companies should consider training models entirely with public domain content or licensed materials, suggested University of Colorado Boulder Associate Professor Casey Fiesler. Still, she acknowledged that some companies might argue that being limited in that way would make their tools weaker.
“Many people consider what’s happening right now to be unethical. But other than that, it’s just sort of wait and see how these copyright cases play out,” Fiesler said. “My guess is that OpenAI is not particularly concerned about these cases.”
Atiya Irvin-Mitchell is a Pittsburgh-based freelance writer and can be reached at airvinmitchell@gmail.com and you can follow her on Twitter/X @atiyawrites.
This story was fact-checked by Sophia Levin.
MOST READ STORIES
Republish our articles for free, online or in print, under a Creative Commons license.
