Paul Tremblay and Mona Awad, two authors, claimed in the lawsuit that OpenAI generated "very accurate summaries" of their works. They believe that these summaries could only be produced by OpenAI after training on their books, which violates copyright law.
On Wednesday, July 5th, according to media reports, two authors filed a lawsuit against OpenAI last week, claiming that their copyrighted books were used to train OpenAI without their consent. This is the second lawsuit that OpenAI has faced within a week due to copyright or privacy issues.
The authors, Paul Tremblay and Mona Awad, stated in the complaint that OpenAI generated "very accurate summaries" of their works. They claimed:
These summaries could only have been produced by OpenAI after training on their books, which violates copyright law.
Paul Tremblay is the author of the book "The Cabin at the End of the World," and Mona Awad's notable works include "Bunny" and "13 Ways of Looking at a Fat Girl."
Introduction to the book "The Cabin at the End of the World" by OpenAI
OpenAI, the globally popular AI chatbot, has undergone extensive training with text data. However, OpenAI has never disclosed the specific data and text used to train its AI, but the company has stated that the chatbot typically scrapes web information, including archived books and Wikipedia.
The lawsuit was filed by the plaintiffs in the United States District Court for the Northern District of California, accusing OpenAI of using "substantial" copyrighted material in its training data, including materials from Tremblay and Awad. However, proving accurately how and where OpenAI collected this information, as well as whether the authors suffered any economic losses as a result, may present challenges for the plaintiffs.
The complaint cites book summaries generated by OpenAI and points out some errors made by the chatbot. However, Awad and Tremblay state that most of the summaries are accurate, indicating that "OpenAI retains knowledge of specific works in its training dataset."
The complaint states:
OpenAI has never copied any copyright management information included in the plaintiffs' published works.
This lawsuit is not the first one that OpenAI has faced.
Last week, a group of anonymous individuals filed a class-action lawsuit alleging that OpenAI's developer, OpenAI, is stealing "a substantial amount" of personal information to train its AI models in pursuit of profit. In this 157-page complaint, the anonymous individuals accuse OpenAI of secretly scraping 300 billion words from the internet, eavesdropping on "books, articles, websites, and posts, including unauthorized access to personal information," in violation of privacy laws.
The law firm Clarkson Law Firm, in a complaint filed last Wednesday with the United States District Court for the Northern District of California, stated that millions of individuals are estimated to be affected by this, and they are seeking potential damages of $3 billion.