Google accused of voice cloning

- The lawsuit was filed in Santa Clara County by David Greene, who hosted NPR's "Morning Edition" for over a decade. He alleges that not just the sound, but also the rhythm and intonation of the NotebookLM voice are so similar that former colleagues asked if he had licensed his voice to Google. - Google spokesperson José Castañeda has denied the claims, stating the voice was created using a paid professional actor and any resemblance to Greene is coincidental. An AI forensics firm hired by Greene's lawyers found a 53% to 60% confidence score that his voice was used to train the AI model. - This is not the first high-profile voice cloning dispute; in 2025, Scarlett Johansson accused OpenAI of imitating her voice, which resulted in the company removing the "Sky" voice option from ChatGPT. Greene's case, however, is the first such dispute to be formally litigated in court. - The case highlights a legal gray area, as U.S. copyright law requires human authorship for protection, a standard purely AI-generated content does not meet according to the U.S. Copyright Office. While California passed a law in 2024 to protect digital replicas, federal legislation has not kept pace. - The NotebookLM tool at the center of the dispute is designed as a research assistant that can generate conversational audio, like a podcast, from user-uploaded documents and notes. - The incident is part of a larger conflict between media creators and AI developers. Several news and media companies, including The New York Times and Intercept Media Inc., have filed lawsuits against AI companies for using copyrighted material to train their models without compensation or permission. - Google's privacy policy explicitly states that it uses publicly available information from the internet to train its AI models, a practice that has raised broader privacy and data rights concerns.