No kidding: Comedian and author Sarah Silverman is one of the lead plaintiffs in a pair of lawsuits against Meta and OpenAI accusing tech companies of illegally using copyrighted works to train their intelligence systems artificial.
Books cited in the lawsuits include Silverman’s best-selling 2010 memoir “The Bedwetter: Stories of Courage, Redemption and Pee.” The federal lawsuits, filed Friday, July 7, allege that OpenAI’s ChatGPT and Meta’s LLaMA both ingested text from “The Bedwetter” and other works to train their large language models (LLMs) — without consent ( or compensation) from authors such as Silverman.
More Variety
Meta declined to comment. OpenAI did not respond to Varietyrequest for comment.
Silverman is one of three authors named as plaintiffs, alongside novelist Christopher Golden (whose books include “Ararat”) and Richard Kadrey, author of the supernatural noir series Sandman Slim. The lawsuits — filed in the U.S. District Court for the Northern District of California, San Francisco Division — seek class-action status and unspecified damages. A copy of the lawsuit against Meta is at this link and the lawsuit against OpenAI is at this link. Lawyers representing the three authors, Joseph Saveri and Matthew Butterick, filed a similar lawsuit against OpenAI last month on behalf of authors Paul Tremblay and Mona Awad.
OpenAI introduced ChatGPT in November 2022. San Francisco-based OpenAI is a private research lab that develops AI technologies, founded in 2015 as a nonprofit by Elon Musk (who no longer sits on the OpenAI board) and CEO Sam Altman.
Although OpenAI did not specify what is included in its datasets for ChatGPT, the lawsuit against the company alleged that the only “Internet book corpuses” that have ever included the volume of material believed to be used by OpenAI are “obviously illegal shadow libraries”. (which allegedly contain the plaintiffs’ copyrighted work). The complaint alleged that when ChatGPT “was asked to summarize the books written by each of the complainants, it generated highly accurate summaries…meaning that ChatGPT retains knowledge of particular works in the training dataset. and is capable of producing similar textual content. At no time has ChatGPT reproduced the copyright management information that the applicants included in their published works.”
In their lawsuit against Meta, plaintiffs’ attorneys alleged that to train the Large Language Model Meta AI (LLaMA) language models, the company copied a massive book dataset that includes the works of the three named authors.
Silverman is a two-time Emmy-winning comedian, actor, writer and producer. In the spring of 2022, her off-Broadway musical adaptation of “The Bedwetter” sold out with the Atlantic Theater Co. She currently hosts “The Sarah Silverman Podcast” and will host TBS’s upcoming “Stupid Pet Tricks,” an offshoot of the David Letterman’s famous late-night segment.
The best of variety
Subscribe to the Variety newsletter. For the latest news, follow us on Facebook, Twitter and Instagram.
Click here to read the full article.