> This is not intended to be AI advocacy

I think it is: It fits the pattern, which seems almost universally used, of turning the aggressor A into the victim and thus the critic C into an aggressor. It also changes the topic (from A's behavior to C's), and puts C on the defensive. Denying / claiming innocence is also a very common tactic.

> You can easily get death threats if you're associating yourself with AI publicly.

What differentiates serious claims from more of the above and from Internet stuff is evidence. Is there some evidence somewhere of that?