Authors Paul Gaetan Tremblay and Mona Awad filed suit on June 28, 2023, against  Open AI, Inc. for training Open AI’s Chat GPT on copyrighted works. The civil case (4:23-cv-03223-AMO) was filed in the United States District Court Northern District of California San Franciso Division. The authors allege six counts of wrongdoing, including copyright infringement, unfair competition, negligence, and unjust enrichment of their work. The pair are also seeking class-action status for a broad class of writers of copyrighted material.

The case could be the beginning of a wave of legal battles against the burgeoning AI industry over issues of privacy and copyright.

Tremblay is an American author and editor of horror, dark fantasy, and science fiction. Mona Awad is a Canadian novelist and short-story writer known for works of darkly comic fiction.

In the filing, the Plaintiffs charge that Chat GPT, along with other large language models, was trained on a collection of more than 7,000 copyrighted books scraped from Smashwords illegally and without permission, along with more than 300,000 long-form works thought to be extracted from “shadow library” websites, such as  Anna’s ArchiveLibrary GenesisSci-Hub and Z-Library. D  Such shadow sites are often hosted without the consent of the original owners of the material.

Filing documents charge that Open AI-Language Models “relied on harvesting massive quantities of textual material from the Internet, including the Plaintiff’s books which are available in digital format.” Open AI made copies of the Plaintiff’s books during the training process of Language Models without permission in direct copyright infringement.

In charging Vicarious Copyright Infringement, the Plaintiffs said: “Because the output of the Open AI language models is based on expressive information extracted from Plaintiffs’ works and others. Every output of the Ppen AI language model is infringing the derivative work made without Plaintiffs’ permission and in violation of their exclusive rights under the copyright act.”

The Plaintiffs charge that when ChatGPT returns summaries from user prompts, the system essentially is creating a “derivative work.” The term is among the ways the Plaintiffs seek to show significant financial to themselves and other writers.

In illustrating how ChatGPT infringes on the Plantiffs’ work, they asked ChatGPT to summarize in detail certain points of one of the Plantiff’s novels, “The Cabin at the End of the World by Paul Tremblay.” A court document then shows CHAT GPT’s complete summary of each piece of the book. It also illustrates a similar scenario for “13 Ways for Looking at a Fat Girl,” by Mona Awad.

The Plaintiffs assert that Open AI knowingly and secretly trained Chat GPT for the company’s own commercial profit without Plaintiff’s or Class’s authorization.

The record says Open AI breached its duties by “negligently, carelessly, and recklessly collecting, maintaining, and controlling  Plaintiff’s and Class members’ infringed Works.”

The Plaintiffs are demanding that the action proceed as a class action with the Plaintiffs serving as Cass Representatives and with the Plaintiffs’ counsel as Class Counsel. The authors are also demanding a trial by jury of all claims in the filing.

The case is being handled by Joseph Saveri Law Firm, LLP in San Franciso.  Matthew Butterick (State Bar No. 250953) of Los Angeles, is serving as Counsel for Individual and Representative Plaintiffs and the Proposed Class,

Meanwhile, the Author’s Guild of America and many writers’ organizations are coming out with statements in an effort to limit Artificial Intelligence in the use of creative material such as books, movies, and graphic creation.

The Guild is introducing four new model clauses concerning AI to its Model Trade Book Contract and Model Literary Translation Contract. The organization reports:

“In addition to the recent clause preventing the use of books in training generative AI without an author’s express permission, the new clauses require an author’s written consent for a publisher to use AI-generated book translations, audiobook narration, or cover art. These clauses can benefit publishers and the publishing industry at large by maintaining the high quality craftsmanship that consumers are used to.

The Authors Guild also urges publishers to identify any books that contain a significant amount of AI-generated text. This summer, the Guild will be publishing AI guidelines for authors and publishers containing each of these conditions.

The purpose of these demands is to prevent the use of AI to replace human creators. The Authors Guild strongly believes that human writing, narration, and translation are vastly superior to their AI mimics.”

Stay tuned for more about the Open AI case.