Meta's Llama model has memorized Harry Potter and Watch Serve the People Onlinethe Sorcerer's Stoneso well that it can reproduce verbatim excerpts from 42 percent of the book, according to a new study.
Researchers from Stanford, Cornell, and West Virginia University analyzed dozens of books from the now-infamous Books3 dataset, a collection of pirated books used to train Meta's Llama models. Books3 is also at the center of a copyright infringement lawsuit against Meta,Kadrey v. Meta Platforms, Inc.The study's authors say their findings could have major implications for AI companies facing similar lawsuits.
According to the research paper, the Llama 3.1 model "memorizes some books, like Harry Potterand 1984, almost entirely." Specifically, the study found that Llama 3.1 has memorized 42 percent of the first Harry Potter book so well that it can reproduce verbatim excerpts at least 50 percent of the time. Overall, Llama 3.1 could reproduce excerpts from 91 percent of the book, though not as consistently.
"The extent of verbatim memorization of books from the Books3 dataset is more significant than previously described," said the paper. But the researchers also discovered that "memorization varies widely from model to model and from book to book within each model, as well as varying in different parts of individual books." For example, the study estimated that Llama 3.1 only memorized 0.13 percent of Sandman Slimby Richard Kadrey, one of the lead plaintiffs in the class action copyright suit against Meta.
So, while some of the paper's findings seem damning, don't call it a smoking gun for plaintiffs in AI copyright infringement cases.
"These results give everyone in the AI copyright debate something to latch on to," wrote journalist Timothy B. Lee in his Understanding AI newsletter. "Divergent results like these could cast doubt on whether it makes sense to lump J.K. Rowling, Richard Kadrey, and thousands of other authors together in a single mass lawsuit. And that could work in Meta’s favor, since most authors lack the resources to file individual lawsuits."
Why is Llama able to reproduce some books more than others? "I suspect that the difference is because Harry Potter is a much more famous book. It's widely quoted and I'm sure that substantial excerpts from it on third-party websites found their way into the training data on the web," said James Grimmelmann, a professor of digital and information law at Cornell University, who was cited in the paper.
What this also shows, Grimmelmann said, is that "AI companies can make choices that increase or reduce memorization. It's not an inevitable feature of AI; they have control over it."
Meta and other AI companies have argued that using copyrighted works to train their models is protected under fair use, a complex legal doctrine. However, the extent of memorization could complicate those arguments.
“Yes, I do think that the likelihood that LLMs are memorizing more than previously thought changes the copyright analysis,” Robert Brauneis, a professor with the George Washington University Law School, said in an email to Mashable. He concluded that the study’s findings could ultimately weaken Meta’s fair use argument.
We asked Meta for comment on the study's findings, and we'll update this article if we receive a response.
Disclosure: Ziff Davis, Mashable’s parent company, in April filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.
Topics Artificial Intelligence Meta
Love, Actually
Huawei previews Nova Flip phone in video, launch set for August 5 · TechNode
Fat bear thieves steal fish from very dominant bear
How to watch the NASA live stream of James Webb telescope images
Right to Burn
The first Webb telescope image is warped and psychedelic. Here's why.
Taobao loosens “refund
With 'Apple Intelligence,' Apple enters the AI era on its own terms
Spielberg’s Children
'Boys Go to Jupiter' review: An animated gem about oranges, aliens, and the gig economy
DACApocalypse Now
Everything AI allows you to do in Windows 11 that saves you time
Blizzard's Hearthstone to return to China on September 25 · TechNode
Charli xcx's 'brat' turns the internet lime green
The Man Who Knew Nothing at All
JBL speakers are up to 38% off at Amazon
James Webb telescope snaps thrilling images of Jupiter and hurtling asteroids
Artificial
We Don’t Have Elections
NASA reveals deep space photo from James Webb Space Telescope test
Staff Picks: Family, Fleece, and Flamin’ Hot Cheetos by The Paris ReviewSpooky Staff Picks by The Paris ReviewThe Nobel Prize Was Made for Olga Tokarczuk by Jennifer CroftThe Charming, Ridiculous Romance Comics of Ogden Whitney by Liana FinckWhat Susan Sontag Saw by Benjamin MoserIn Russia, the Ultimate Scary Story is about Losing Your Coat by Jennifer WilsonVoyage around My Cell by Ahmet AltanWho Was My Mother? by Sallie TisdaleThe Many Reincarnations of Kim Deitch by Bill KartalopoulosRedux: Tautology, Tautology by The Paris ReviewThe Enigma of Prince: An Interview with Dan Piepenbring by Cornelia ChanningCooking with Shirley Jackson by Valerie StiversA Polyphonic Novel of Midcentury San Francisco by Jessica HagedornDoes Poetry Have Street Cred? by Major JacksonConsider the Butt by Heather RadkeThe Jets, the Bills, and the Art of Losing by Rowan Ricardo PhillipsStaff Picks: Metaphors, Messengers, and Melancholy by The Paris ReviewRedux: The Deep Well of Other Beings by The Paris ReviewHarold Bloom’s Immortality by Lucas ZwirnerPoetry Rx: The Radiant Bodies of the Dead by Claire Schwartz Guess which gender is more likely to be verified on Twitter? We'll wait. Watch this clever rugby dude pull off a stunt 15 years in the making Snapchat inks deal with Foursquare for more precise geofilters Pilot goes on epic rant after political fight breaks out on airplane #EarthToMarrakech: COP22's digital call Thousands of dead fish fill Long Island canal Your guide to prepping for 'Fantastic Beasts and Where to Find Them' 'Super Mario Run' comes to the iOS App Store in December Kids riding on trains for the video views aren't thrilling cops China, India would pick up America's clean energy slack under Trump Sombra has arrived in 'Overwatch' Miss Universe contestant responds to accusations of racism The power of public shaming in Trump's America Wordle today: The answer and hints for June 18, 2025 Grumpy man interrupts TV interview to rant about meddling young people Trump gets ghosted by New Zealand PM because earthquakes will do that Klay Thompson to the Celtics? How fake news infiltrates sports, too. A Samsung Galaxy S7 phone exploded in Canada NYC taxi driver pin The Flash Season 3, Episode 6 sees Wally face off with Alchemy
2.2584s , 10130.7109375 kb
Copyright © 2025 Powered by 【Watch Serve the People Online】,Exquisite Information Network