There are hundreds of OpenAI employees. Anyone of which is capable of testifying to the extent of copyrighted material present in the training datasets. It does

There are hundreds of OpenAI employees. Anyone of which is capable of testifying to the extent of copyrighted material present in the training datasets. It does not take too much work to get any number of agents to regurgitate extensive verbatim quotes from books/movies/newspaper articles/whatever.