Combined with Murati’s reaction when asked if they trained Sora on YouTube videos, it’s obvious that OpenAI has trained their TTS systems on a whole bunch of copyrighted content including the output of professional actors and voice actors who definitely weren’t compensated for their work.
Altman and Murati are world-class grifters but until now they were stealing from print media and digital artists. Now they’re clashing with some of the most litigious industries with the deepest pockets. They’re not going to win this one.