We can't expect people to somehow deduce that AI CEOs are lying and therefore should like them.
Also, even if people DO know AI CEOs are lying... why would that make them likable? Liars are unlikable, almost as a rule. Of course people don't like and trust them.
So either AI CEOs genuinely believe the "permanent underclass" stuff, in which case you should hate them because obviously. Or they don't believe it, in which case you should hate them because they're liars. Either way they lose. The only way to win is to not say batshit insane stuff that implies everyday people will die en-masse. Which... yeah. I feel like anyone could've told them that.