Nvidia is being sued by three authors. They accuse the chip company of using their copyrighted books to train the AI framework NeMo – without permission. According to Reuters, the authors also explain that their books were only part of a data set of 196,640 books that were used as training material in NeMo. They are therefore demanding compensation for everyone in the USA who is affected.
Advertisement
Although the NeMo framework is still available for developers via Github, among other things, the cloud service has no longer been able to deliver results since October 2023, writes Reuters – allegedly due to “reported copyright infringements”.
However, NeMo will no longer be available since October 2023, writes Reuters – allegedly due to “reported copyright infringements”. In the lawsuit, which was filed in federal court in San Francisco, the authors claim that by shutting down NeMos, Nvidia essentially admitted that it had violated copyrights. The plaintiffs' works are Brian Keene's 2008 novel “Ghost Walk,” Abdi Nazemian's 2019 “Like a Love Story” and Stewart O'Nan's 2007 novella “Last Night at the Lobster.”
Copyright lawsuits against AI providers
Nvidia has not yet commented on the allegations. The lawsuit joins other cases in which authors accuse providers of language models of having used their works unlawfully. The New York Times lawsuit against OpenAI has attracted the most attention so far. They are demanding several billion US dollars in damages because, as they say, OpenAI is supposed to make a lot of money with its language model – which is only possible because they used copyrighted material for training. And OpenAI actually said that it wouldn't work without such training material. That's why the AI company relies on the American special regulation of “fair use”. This allows exceptions if it benefits everyone. It is still questionable whether a commercially sold tool can benefit everyone.
In addition to damages, a decision in favor of the New York Times could also mean that the copyrighted content would have to be deleted from the language models. That would have far-reaching consequences for ChatGPT and Co.
Nvidia has become the darling of the stock market in recent months. All AI companies also use the manufacturer's chips. Companies like Microsoft and Google are literally ripping the accelerators for training AI algorithms out of Nvidia’s hands.
Aktenzeichen: Nazemian et al v Nvidia Corp, U.S. District Court, Northern District of California, No. 24-01454
(emw)