OpenAI and Google trained their AI models on text transcribed from YouTube videos, potentially violating creators’ copyrights, according to The New York Times. The report, which describes the lengths ...