I already left one thought on this previously. Not sure if it will be read. But now that I've seen the AI voice option in practice it's even worse than I thought it would be. Seriously? It reads everything on the page, which means it reads the captions/credits for the photos, which sounds ridiculous...if the writer has done a pull quote, for visual appeal (as we're encouraged to do), it reads that, even though the pull quote is typically something that already appears in the actual text or will after it appears as a pull quote, which means it will be read twice and won't make sense to the listener the way it does visually to the reader...this is absurd. If you let authors voice their own work we would know better, but your coders either didn't think about this (of course, they're programmers, not writers or readers), or they did and just didn't care...seriously, either let people voice our own pieces or get rid of this...or at least let writers turn OFF this option for our own pieces if we can't voice them.