A trio of authors filed copyright infringement lawsuits against Meta and OpenAI, the ChatGPT maker, for allegedly using their material to train chat bots. Image source: Gage Skidmore/Flickr
US comedian Sarah Silverman and two authors have filed lawsuits against Meta Platforms and OpenAI, the maker of ChatGPT, for allegedly training their chat bots using copyrighted material.
The proposed class action lawsuits filed in San Francisco federal court last week by Christopher Golden, Richard Kadrey, and Silverman allege that Meta (parent company of Facebook) and Open AI used their content without permission to train AI language models.
These lawsuits highlight the legal risks that chat bot developers face when using hoards of potentially copyrighted content to develop apps that create realistic responses from user prompts.
Silverman, Golden and Kadrey all allege that OpenAI and Meta used their books without permission to develop their language models to replicate human conversation.
In the trios’ lawsuit against Meta, it os alleged that leaked information regarding Meta’s AI business shows their work was accessible in datasets used by Meta without authorisation to train its LLaMA models—an array of open-source AI models.
The lawsuit against OpenAI alleges that ChatGPT generated summaries of the writers’ works which indicate that their work was included in the bot’s training.
The lawsuit against OpenAI alleges that ChatGPT generated summaries of the writers’ works which indicate that their work was included in the bot’s training.
According to the lawsuit, ChatGPT did get some details wrong in its summaries, but “retains knowledge of particular works in the training dataset.”
In both lawsuits, the trio of authors state that they “did not consent to the use of their copyrighted books as training material” for both companies chat bots.
On their website, the lawyers representing the writers, Joseph Saveri and Matthew Butterick, state that they have heard from “writers, authors, and publishers who are concerned about [ChatGPT’s] uncanny ability to generate text similar to that found in copyrighted textual materials, including thousands of books.”
On behalf of a nationwide class of copyright owners, the lawsuit is seeking unspecified monetary damages for the worked that were allegedly infringed.
