New state assembly bill regulates, provides transparency surrounding generative AI
Get the latest Syracuse news delivered right to your inbox.
Subscribe to our newsletter here.
New York State Assemblymember Jonathan Rivera introduced a bill in September to mitigate the use of artificial intelligence-generated content in online book selling.
The bill, proposed on Sept. 29, would require online booksellers who are selling content created wholly or partially with generative AI to disclose such use before a sale.
“There’s no responsibility at all to have these people that are just churning out what they’re calling books and selling them quickly purely to make a profit, not to educate children, not to inform people, not to bring well-thought-out literature to people,” Rivera said. “It’s all just a money grab.”
Rivera, who represents New York’s 149th District, encompassing parts of Buffalo and Lackawanna, said he was motivated to propose the bill after hearing about a scam class that taught people how to make money quickly by mass producing AI-generated books on Amazon.
“The problem was that the end product in all of these examples was that the actual book, coloring book, novel, whatever it was, was so poorly put together and subpar,” Rivera said.
Online sellers of AI-generated books often use AI as a pathway to generate income regardless of the impact on the user, Rivera said. He said these books often disproportionately target children, citing children’s short stories void of structure and coloring books containing distorted images and confusing anatomical inaccuracies.
“As this new technology becomes interwoven into the publishing world, consumers deserve full transparency about who (or what) wrote, illustrated, or otherwise designed the books they are buying for themselves, their children, their students, or new language learners,” the bill’s justification section states.
Cindy Zhang | Digital Design Director
Shubha Ghosh, director of SU’s Intellectual Property Law Institute and Crandall Melvin Professor of Law, said Rivera’s proposed bill represents an effort to address potential risks that generative AI poses to consumers.
“This is sort of a first step in terms of trying to regulate and deal with some of the perceived dangers of using AI technology and publishing,” Ghosh said.
According to an article by Axios, searches on Amazon’s website, which is estimated to control at least half of U.S. book sales and a large share of the growing e-book market, are increasingly producing AI-generated titles filled with “unreliable information and soggy prose.”
As of September 2023, Amazon now requires some authors to disclose when content being sold is AI-generated; however, authors do not have to disclose if the AI was only used to “brainstorm and generate ideas.”
Rivera said this bill will help keep people informed about the origin of the content they are purchasing and consuming.
“If this bill passes, the (consumers) would be able to see clearly at the point of purchase that it was AI-created and it wasn’t created by somebody that uses a filter and a brain and their own creativity behind it,” Rivera said.
Catherine Cocks, director of the Syracuse University Press, said while she’s not opposed to the bill, she isn’t convinced it will achieve the intended purpose of keeping consumers informed.
“Readers should know if they’re reading the results of another person’s creative insights and hard-won writing skills, or a remixing of people’s creative work by software,” Cocks wrote in an email statement to The Daily Orange. “But simply making readers aware may not have much effect on how books get written and published.”
Ghosh acknowledged the bill will serve an “educational function” but said it needs more detail on what exactly to disclose about books’ use of AI.
“It all depends on how people understand the disclosure and whether they think about the harm from it,” Ghosh said.
Cocks also pointed out questions that the bill leaves unanswered, such as what it means to say a book was “wholly or partially” created by generative AI and if the disclosure requirement includes books that use AI-assisted research.
“Simply disclosing the use of AI in the creation of a book might not actually tell the reader much about the content of the book or the circumstances of its production,” Cocks told The D.O. “We have a lot of thinking to do to figure out the best way to manage generative AI for the public good.”
Ghosh also warned that lawmakers should be wary of reactive legislation until there’s a clearer general understanding on AI technologies’ impacts.
“I’d be wary of just reactive legislation at this point, before we really understand what the impact would be,” Ghosh said.
Despite these uncertainties, Cocks and Rivera both recognized this bill is an important step towards protecting the work of authors and other creatives from the potential risks of generative artificial intelligence.
“They’re confronting a technology that is looking to put their likeness and their creativity out the window,” Rivera said. “You have huge corporations looking to profit off of someone else’s work that they can just sort of re-engineer to make their own without the permission or support of the original author of that creative work.”
Now that the bill has been proposed, a state senator must sponsor a companion bill to be reviewed in both the Assembly and Senate when session begins in January 2024, Rivera said. If passed, Gov. Kathy Hochul would sign the bill into state law.
Rivera believes other representatives need to do more to address the rising presence of artificial intelligence in every industry beyond publishing nationwide.
“We could all do more because the technology has so many ramifications in so many industries,” Rivera said. “We as representatives need to hear from industries all over the country to let us know how it’s affecting them because we’re sort of still learning as we go.”
Cocks expressed a similar sentiment that people in all industries need to leverage the advantage of generative AI.
“New technologies like generative AI can have tremendous beneficial effects if we design and implement them with the well-being of people and the planet in mind,” Cocks wrote. “If we don’t think carefully and act thoughtfully, such technologies can have terrible effects.”
