Nvidia is facing a lawsuit filed by a trio of writers who claim that the chipmaker trained its NeMo artificial intelligence (AI) platform using their copyrighted novels without their consent.
Alleged Copyright Infringement
According to Reuters, the works of Brian Keene, Abdi Nazemian, and Stewart O'Nan were removed from the NeMo dataset last October "due to reported copyright infringement." The dataset included around 196,640 volumes that were used to train NeMo to mimic normal written language.
On Friday night, March 8, the authors argued in a federal lawsuit filed in San Francisco that the removal shows that Nvidia has "admitted" it trained NeMo on the dataset, violating their copyrights. They are requesting an undisclosed amount of money for Americans whose copyrighted works were used in the previous three years to train NeMo's large language models.
This case encompasses a number of literary works, including Ghost Walk by Keene (2008), Like a Love Story by Nazemian (2019), and Last Night at the Lobster by O'Nan (2007).
Growing Trend but With Downsides
Due to the increasing popularity of AI, Nvidia has been a darling among investors.
Nvidia is a chipmaker situated in Santa Clara, California. Its stock price has increased by about 600% since the end of 2022, valuing the company at nearly $2.2 trillion.
However, this recent lawsuit adds to the increasing number of lawsuits filed by authors, and even the New York Times writers, concerning generative AI. This sort of AI creates new material by analyzing inputs, including text, photos, and audio. NeMo is an easy and inexpensive technique for implementing generative AI, according to Nvidia.
Companies like OpenAI and its partner Microsoft, who developed the AI platform ChatGPT, have also been sued over the tech.
