The Authors Guild, a U.S. trade group for writers, filed the proposed class-action on Tuesday on behalf of 17 plaintiffs, including big literary names John Grisham, Jodi Picoult, Roxane Gay, Suzanne Collins, Doug Preston George R.R. Martin and Jonathan Franzen.
The Authors Guild’s proposed class action says writers should be compensated for the use of their work
Sheena Goodyear · CBC Radio
As It Happens5:59Authors launch lawsuit accusing Open AI of pirating their books to train ChatGPT
When novelist Douglas Preston first saw ChatGPT, he decided to do a little experiment.
He asked the artificial intelligence-powered chatbot to write a poem about A.X.L. Pendergast, a fictional character from a crime novel series he writes with Lincoln Child.
“It wrote this marvellous poem in heroic couplets. But when I read it, I realized that it knew so much about this character — way more than it could possibly have known just from scraping Wikipedia or reviews of the book,” Preston told As It Happens host Nil Köksal.
“That’s when I got a little bit disturbed and I thought: It ingested my books.”
Preston is now one of several authors who are launching a lawsuit against Open AI, the maker of ChatGPT, accusing the company of illegally pirating hundreds of books online and using them to train its AI without consent or compensation.
The Authors Guild, a U.S. trade group for writers, filed the proposed class-action on Tuesday on behalf of 17 plaintiffs, including Preston, George R.R. Martin, Jodi Picoult, Michael Connelly and Jonathan Franzen.
OpenAI did not respond to the allegations of piracy, but in a statement to CBC, said: “We respect the rights of writers and authors, and believe they should benefit from AI technology.
“We’re having productive conversations with many creators around the world, including the Authors Guild, and have been working co-operatively to understand and discuss their concerns about AI,” the statement reads.
“We’re optimistic we will continue to find mutually beneficial ways to work together to help people utilize new technology in a rich content ecosystem.”
‘We are here to fight’
AIs like ChatGPT learn to generate new content by first ingesting huge swaths of existing content — whether it be art, text, or music — often scraped from the internet.
Whether that constitutes a violation of copyright law remains to be seen, but it’s a question that’s being tested in court.
Several companies that use generative AI — including Meta Platforms and Stability AI — are facing similar lawsuits by writers, visual artists and source-code writers.
Trademark and copyright lawyer Daniel Anthony, of Toronto’s Smart & Biggar, says there could be a case to make that ChatGPT violates copyright law — though it depends very much on the details.
“Let’s assume that a particular AI product was trained on the book Moneyball. How did they acquire a copy of that book? Was it purchased or simply found on the internet (which could even be a pirated copy)? Even if purchased, did the AI ‘read’ the book, or was a fresh copy made and entered into a database for the training?” he told CBC in an email.
“These details may be relevant in determining whether unlawful copying has taken place.”
AI defendants have argued that their use of training data scraped from the internet qualifies as fair use under U.S. copyright law.
The Authors Guild says there’s nothing fair about it.
“To preserve our literature, authors must have the ability to control if and how their works are used by generative AI,” Maya Shanbhag Lang, president of the Authors Guild, said in a press release.
“Our membership is diverse and passionate. Our staff, which includes a formidable legal team, has expertise in copyright law. This is all to say: We do not bring this suit lightly. We are here to fight.”
The Authors Guild is calling on OpenAI and other creators of generative AI to find a way to fairly compensate the creators whose work is being used to train their systems.
Already, it says, authors are losing income to AI-generated books being sold on Amazon and passed off as real. In some cases, it says, people have tried to use AI to mimic specific authors.
“For example, one of the plaintiffs, my friend George R.R. Martin, somebody misused ChatGPT to write the final book in his series that he hasn’t written yet,” Preston said.
“And George, you know, told me that he’s really hurt by that and and really disturbed and upset and angry, as well he should be.”
Just last month, author Jane Friedman told As It Happens that she found at least five AI-created books for sale on Amazon with her name on them.
“I was expecting something like this to happen eventually. I just didn’t think I would find myself leading the charge on fraudulent work in my name,” she said at the time.
The lawsuit is calling for the creation of a licensing system that would allow AI companies to use and pay for copyrighted material.
Authors who don’t want their work used for this purpose could opt out.
“We authors just want to be able to control our own work and the destiny of our work. We don’t want these giant tech companies coming in and creating extremely valuable commercial products, basically by stealing our work. It’s as simple as that,” Preston said.
Otherwise, he says: “It’s a little bit like coming home and finding that someone has been in your house and taking things.”
With files from Reuters. Interview with Douglas Preston produced by Chris Trowbridge