Authors sue ChatGPT creators for using their books to train AI without consent

Authors Mona Awad and Paul Tremblay are suing OpenAI for allegedly using their books to train ChatGPT without their consent.

On June 28, two authors filed a lawsuit against OpenAI, the creators of the artificial intelligence chatbot. The lawsuit claims that the company breached copyright laws by “training” their AI dataset with their books without the author’s explicit consent. 

The award-winning authors, Monda Awad and Paul Tremblay, filed the lawsuit claiming that their copyrighted books were “copied and ingested” into the AI’s dataset without their consent, which infringes upon their rights as writers.

The complaint argues that when ChatGPT is prompted, it can generate summaries of both their works, which is only possible if the AI was trained on their copyrighted works, the lawsuit claims.

Because ChatGPT is a software product that is being sold by OpenAI, their usage of Awad’s and Tremblay’s works in training the AI dataset reportedly infringes upon the copyright of their works. As the AI tool is being sold for profit by the company in focus. 

The lawsuit writes, “[OpenAI], by and through the use of ChatGPT, benefit commercial and profit richly from the use of [Monda Awad and Paul Tremblay’s] copyrighted materials.” 

According to IP law expert Andres Guadamuz in a comment to The Guardian’s report on the lawsuit, this is the first lawsuit brought against ChatGPT that concerns copyright. 

The lawsuit comes amid long-standing concerns from creatives as AI tools are able to create texts and images in seconds, all through their machine learning algorithms made through the works of existing pieces of media. 

This has prompted organizations such as The Authors Guild to pen an open letter to companies that sells AI tools to obtain consent, credit, and compensate writers for the use of their copyrighted materials in training their AI.