Comic claims AI tool – and Meta's similar bot – learned from illegally copying her book

Sarah Silverman is suing the companies behind two of the most powerful AI engines for stealing her work.

The comedian claims OpenAI – who created ChatGPT – and Facebook owner Meta breached the copyright of her memoir Bedwetter.

Legal papers show that when asked to summarise the book, the AI engine can go into detail about its content. The lawsuit alleges this shows that the bot was trained on illegally-acquired datasets containing her work, as pirated copies of the book  are available online. 

She is suing with authors Christopher Golden and Richard Kadrey, who are all similarly claiming that Meta’s  LLaMA models also used ‘flagrantly illegal… shadow libraries’ to acquire information.  

In both claims, first reported by tech website The Verge, the authors say that they ‘did not consent to the use of their copyrighted books as training material’ for the AI models and seek unspecified  damages.

Large language AI models scan huge databases of text to learn how to be able to reproduce con­vinc­ing sim­u­la­tions of nat­ural written lan­guage.

But law firm LL Mitigation – which is bringing the claims – say on their website: ‘"Gen­er­a­tive arti­fi­cial intel­li­gence" is just human intel­li­gence, repack­aged and divorced from its cre­ators.

Click here to read the Silverman, Golden and Kadrey, lawsuit in full – while here is the exhibit they have submitted to the court in San Francisco, showing how Chat GPT had a detailed knowledge of Bedwetter. 

Published: 10 Jul 2023