The media conglomerate behind USA Today, along with the owner of several of its affiliated newspapers, has initiated legal proceedings in a New York federal court. They accuse OpenAI of extensively utilizing their copyrighted news articles without permission to train and operate its generative artificial intelligence models, subsequently redistributing this content through platforms like ChatGPT. The plaintiffs are demanding compensation in excess of $250 million, along with additional judicial actions to curb the use of the implicated models.
The complaint details that USA Today and its 13 associated entities claim OpenAI replicated hundreds of thousands of news articles and other copyrighted materials during the development of its GPT series models. These works were incorporated into the datasets employed for model training, allowing the models to assimilate, encode, and eventually leverage the information and expressions from these articles when generating responses for users.
Furthermore, the plaintiffs accuse OpenAI of deliberate infringement, asserting that during the model's development and training phases, OpenAI was, or should have been, aware that the works were under copyright protection, yet proceeded to replicate this content extensively.
Beyond seeking financial restitution, the lawsuit urges the court to mandate the destruction of the GPT models trained with the disputed media content and their associated training datasets. It also calls for a jury trial. The media outlets involved in the lawsuit represent prominent newspapers and news organizations spanning various regions across the United States, totaling 19 publications.