A judge in California dismissed a large portion of a copyright lawsuit that claimed OpenAI used the authors’ books without permission to train its famous ChatGPIT chatbot.
In 2023, comedian Sarah Silverman, writers Michael Chabon and Ta-Nehisi Coates and some other writers filed a lawsuit. They said that OpenAI and its parent company Microsoft unlawfully took and used their copyrighted literary works to enhance ChatGPIT’s natural language capabilities.
On Monday, a US district judge named Araceli Martínez-Olguin issued a ruling that mostly agreed with Microsoft-backed OpenAI’s request to dismiss the lawsuit. For now, at least, the judge was not convinced by arguments that ChatGPT infringes the creative rights of the authors themselves or that OpenAI unfairly makes money from their books.
The bigger problem that has not yet been resolved is whether the AI models generating content are breaking copyright laws by using massive amounts of data from the Internet without permission. Big tech companies argue that this type of use is appropriate and actually promotes important AI advancements that help the general public.
Lawyers representing both the plaintiffs and OpenAI had no immediate comment when contacted Tuesday.
Groups have recently filed lawsuits against OpenAI, Meta Platform, Microsoft, and others, claiming they are using copyrighted songs, art, and media content without permission in their AI training data sets.
Judge Martínez-Olguín sided entirely with OpenAI and agreed that ChatGPT’s specific outputs did not in fact match the authors’ original literary works. According to the present case law standards this completely counts as a violation.
He told the authors they have until March 13, 2024, to fix their complaint and get rid of the significant claims. The authors said this week that AI copyright cases in New York are simply copying each other and should either be combined or thrown out.
The main purpose of this court battle is to find out whether using copyrighted online content without permission or payment is breaking the law and to push the boundaries of AI advancement, especially with advanced chatbots like ChatGPT. It is important to consider the benefits to the public.
For now, this California judge is saying that ChatGPT’s current output doesn’t actually cross that line, even though the way it has been trained may still be legally incomplete. This decision, in a way, supports OpenAI’s argument but does not completely resolve the AI copyright issue.