You do however have the ability to know how offensive it sounds.
That's precisely what I said I don't:
"neither have much ability to know how "offensive" it sounds"
Which means, I wouldn't even know how to fix it, and when it stops being "offensive".
It is very clearly an american accent so even if it didnt use racist stereotypes it is still wrong, and in a very obvious way.
Very obvious to you maybe, as native speaker, and it doesn't sound very american at all to me (and I guess to most americans), considering the voice is cloned from a japanese person speaking english.
I'll make an example, please don't take it again as a "racist" example, and try not to abuse of that world, which seems to be a catch-all buzzword for anything you don't like. I'm an Italian native speaker, and I don't THINK I sound like Super Mario when I speak English. Problem is, to English-native speakers, I probably DO sound like Super Mario, even it I might be fully convinced I'm not, and my italian friends might agree with me. Problem is, is not for us to decide how we sound when we speak English: only native English speakers can say that.
So, instead of accusing an AI of being "racist", while you don't act proactively and try to source a real voice ? Some users DID helped with accents already, wouldn't be the first time.