class=”sc-29f61514-0 icZBHN”>
In a class-action lawsuit filed in a federal court in New York, US writers’ association Authors Guild and other authors such as Martin and John Grisham accuse the California start-up of using their books to train artificial intelligence without their consent. However, these are protected by copyright.
ChatGPT imitates the writers’ style
According to the class-action lawsuit filed Tuesday, the “heart” of the large language models on which ChatGPT is trained relies on “systematic theft on a massive scale.” These language models “endanger the livelihood of writers because they allow anyone to create texts automatically and for free (or so cheaply) that they would have to pay writers.”
With the help of ChatGPT, content that mimics the authors’ style can be created, the plaintiffs said. Intentional copying “perversely” “turns plaintiffs’ works into machines of their own destruction.”
In addition to damages, the plaintiffs are demanding that their copyrighted books not be used to develop algorithms for generative artificial intelligence technology “without express permission.” OpenAI did not immediately respond to AFP’s request for comment.
OpenAI did not disclose training data
The company leveraged mountains of text available online to fuel ChatGPT’s language model. The algorithm allows the chatbot to communicate with users through text-based messages. However, OpenAI has never fully disclosed which websites and fonts are used for this purpose.
(AFP)
Source :Blick
I’m Tim David and I work as an author for 24 Instant News, covering the Market section. With a Bachelor’s Degree in Journalism, my mission is to provide accurate, timely and insightful news coverage that helps our readers stay informed about the latest trends in the market. My writing style is focused on making complex economic topics easy to understand for everyone.
