California lawmakers shouldn’t require artificial intelligence developers to disclose to copyright holders whether their works are in a model’s training data set, a representative for OpenAI told a legislative panel Monday.
AI companies are collaborating more with copyright holders as they build up their models, Mark Gray, copyright policy counsel for OpenAI, told lawmakers. He cited OpenAI’s contract with news organizations such as the Associated Press and its work to display more excerpts of news from original sources on OpenAI’s platform. The New York Times has sued OpenAI for copyright infringement in a closely watched case on the topic.
Lawmakers ...