Sarah Silverman sues OpenAI and Meta. Know the exact reasons

Author and comedian Sarah Silverman is filing a case against OpenAI and Meta along with two other authors, Richard Kadrey and Christopher Golden, in a US District Court over claims of copyright infringement. 

Lawsuit against OpenAI

Among other things, the suits allege that ChatGPT by OpenAI and LLaMA by Meta have been trained on datasets that are illegally acquired, and which contain their works acquired from "shadow libraries" websites such as Library Genesis, Bibliotik, Z-Library, and many more. They say that the books are available in bulk through torrent systems.  When asked, Kadrey and Golden refused to make any comments on the lawsuit. The team of Silverman too did not give an answer.

In the OpenAI lawsuit, the authors say that when prompted, OpenAI's ChatGPT is going to summarize their books and thereby infringe on their copyrights.  Bedwetter by Silverman is the first book that could be seen summarized by ChatGPT. Ararat by Golden and Sandman Slim by Kadrey are also examples. The claims in the lawsuit also say that ChatGPT did not provide any of the copyright management data that the authors included along with their published pieces of works.

Lawsuit against Meta

In the suit against Meta, the plaintiffs allege that their books were actually accessible in datasets used by Meta to train LLaMA models introduced in February.

The suit states in steps the reason why the authors believe the datasets contain illicit origins, i.e. in a Mera paper containing LLaMA. The company states that one of its dataset sources is ThePile.  EleitherAI is the company that assembles ThePile. The lawsuit points out that ThePile was described in the EleutherAI paper as being put together from "a copy of the contents of the Bibliotik private tracker". Bibliotik, along with other shadow libraries listed in the lawsuit are alleged to be flagrantly illegal. 

The claims

The authors say that they "did not consent to the use of their copyrighted books as training material" for the AI models of the companies. The lawsuits each carry six counts of multiple types of copyright violations, unfair competition, unjust enrichment, and negligence. The authors are seeking statutory damages and restitution of profits.

ALSO READ: What is the first Quantum Supercomputer by Microsoft?