Meta Platforms Inc (NASDAQ: META) have gotten themselves into a heap of trouble again.
The social media giant has been accused of illegally using other people’s copyrighted material to train its artificial intelligence models.
The ongoing suit is called Richard Kadrey et al v. Meta Platforms, and it’s spearheaded by a group of high profile authors.
First among them is the Pulitzer Prize winner Andrew Sean Greer and the National Book Award winner Ta-Nehesi Coates. Mark Zuckerberg claims that using over seven million books from the pirate library LibGen falls under “fair use” laws and was therefore legal.
In a new Vanity Fair piece, the writer highlights Meta’s latest legal argument: the books it used to train its multibillion-dollar language models are worthless.
Meta’s attorneys argue that these books offered little value. They claim the content didn’t significantly impact the performance of their AI models. Furthermore, Meta brought in an expert witness to support the claim. He downplayed the influence of any single book on the model’s output.
According to the expert, one book altered the language model’s performance by less than 0.06 per cent on standard benchmarks. He claimed this effect was meaningless—essentially just noise—and not enough to warrant concern over copyright or compensation. This defense attempts to reframe the AI training process as minimally reliant on any specific copyrighted work.
Meta’s argument is that there’s no market in paying authors for their copyrighted works because there’s nothing of value to exchange. Evidently, none of the author’s works carries any economic value as training data—and therefore, there’s no breach of copyright.
Read more: Verses AI anticipates launch of artificial intelligence program in April
Read more: Fetch.ai builds bridge between Web3 and artificial intelligence
Creative theft no longer a problem because AI writes as well as humans
This is typical of the sleight of hand and double-speak Meta—and much of the AI industry—uses when challenged on content use. They insist human-created material isn’t that valuable, so everyone should stop clutching their pearls over the sanctity of art.
At the same time, they claim AI writes just as well as humans now—no need to worry about creative theft. But then they argue that this same content is absolutely vital to building AI models that will save the world. So, they say, don’t make them pay for using it.
That particular stretch of yarn came from AI mavens, OpenAI, themselves earlier this year. They told the British Parliament that the public domain doesn’t offer enough material to train its models to the required level. To compensate, it said it must freely use contemporary copyrighted works without paying creators anything.
This seems to reflect an unspoken agreement among top AI companies. When a Meta researcher asked if the legal team had approved using LibGen, another replied: “I didn’t ask questions, but this is what OpenAI does with GPT-3, what Google does with PALM, and what DeepMind does with Chinchilla—so we’ll do it too,” according to Vanity Fair, citing internal messages from the lawsuit.
“In no case would we disclose publicly that we had trained on LibGen, however there is practical risk external parties could deduce our use of this dataset,” an internal Meta slide deck read.
Read more: Google creates AI model to help understand what dolphins are talking about
Read more: Meta Platforms to add its own standalone artificial intelligence app
Zuckerberg can’t keep fingers out of cookie car
This isn’t the first time Mark Zuckerberg and Meta did an end-run around common decency to achieve its aims, either. In fact, he’s practically famous for it.
Over the past decade, Meta has faced repeated scrutiny over serious ethical breaches, particularly concerning user privacy, content moderation, and political manipulation.
In 2018, whistleblowers and journalists revealed that Cambridge Analytica had harvested the personal data of up to 87 million Facebook users without their consent. This data helped build psychological profiles to target voters during the 2016 U.S. presidential election and the Brexit campaign, raising alarm about Facebook’s role in undermining democracy.
Beyond data misuse, Meta has consistently struggled to police harmful content on its platform. Whistleblower Frances Haugen testified in 2021 that Facebook prioritized engagement over safety, allowing hate speech and misinformation to spread unchecked.
Internal research showed that Instagram negatively affected teen mental health, especially among girls, yet the company failed to act meaningfully.
Critics have accused Meta of inconsistently enforcing its content policies and bowing to political pressure. In some cases, the company allowed inciting content to remain online, contributing to real-world violence. Despite public statements about improving transparency and safety, critics argue that Meta continues to place profits above ethical responsibility, with little meaningful change to its core practices.
.
