Title: Lawsuits Threaten ChatGPT and Meta’s AI Engines Over Copyright Infringements
A wave of lawsuits filed by authors against OpenAI, the owner of ChatGPT, and Meta, the developer of LLaMA, could potentially lead to significant changes in how their AI engines handle copyright material. Failing to adapt or compensate the authors may result in OpenAI and Meta having to pay out millions in damages.
OpenAI’s ChatGPT gained early market recognition by ingesting a large amount of online content without sufficient consideration for copyright ownership. This approach helped OpenAI establish a prominent position in the market and attract significant investments. However, this strategy now poses a legal challenge as the publishing community begins to contest these actions in court.
Meta’s LLaMA, a 65-billion parameter language model released to compete with ChatGPT, is also under scrutiny. The publishing community associates both ChatGPT and LLaMA with industrial-strength plagiarists that violate the rights of book authors, according to attorneys Joseph Saveri and Matthew Butterick.
In response to copyright breaches allegedly committed by OpenAI, Saveri and Butterick have filed a class-action lawsuit on behalf of authors Sarah Silverman, Chris Golden, and Richard Kadrey. This lawsuit follows a similar class-action filed last month on behalf of authors Paul Tremblay and Mona Awad. The attorneys have also initiated legal action against the AI image generator Stable Diffusion.
The litigation challenges the courts to determine whether AI engines can freely ingest copyrighted material published online. Denying the existence of copyright breaches would effectively sanction the unauthorized use of personally owned copyrighted content.
Saveri and Butterick argue that ChatGPT’s training relied on text summaries generated from copyrighted works. They claim that OpenAI and Meta sourced much of their training data from copyrighted books without proper consent, acknowledgement, or compensation.
Books, known for their value as training data, have a distinctive advantage within the AI community. The significance of books as training material was even highlighted by researchers from MIT and Cornell, who found that books had the greatest positive impact on machine learning performance.
The plaintiffs in the lawsuit include highly successful and reputable authors. Sarah Silverman is a two-time Emmy Award-winning comedian, actress, and author of the best-selling memoir The Bedwetter: Stories of Courage, Redemption, and Pee. Christopher Golden is a New York Times bestselling author and recipient of the Bram Stoker Award for his works such as Road of Bones, Ararat, Snowblind, and Red Hands. Richard Kadrey is also a New York Times bestselling author known for his supernatural noir series Sandman Slim. Paul Tremblay, the author of crime novels like The Little Sleep and No Sleep Till Wonderland, and Mona Awad, whose novel Bunny was named a Best Book of 2019 by Time, Vogue, and the New York Public Library, are among the plaintiffs as well.
These lawsuits have been filed in the US District Court in San Francisco, with the authors seeking damages and a civil trial by jury. The outcome of these legal battles could have significant implications for the future of AI engines and their use of copyright material. OpenAI and Meta may be forced to adapt their practices or face substantial financial consequences.
As these cases progress, the courts will ultimately decide the fate of AI engines like ChatGPT and Meta’s LLaMA. In the ever-evolving landscape of artificial intelligence, striking the balance between innovation and respecting intellectual property rights remains a critical challenge.
