You are so right about that!

The desire to quietly move people into a fictional bubble to "neutralize" them, of managing people without them knowing it, and keeping them busy so they don't do something you don't want them to do. It's like having a physically unfit student demand very aggressively to be put on the football team, and instead of telling him he doesn't make the cut because you're afraid he might pull out a gun, you put some VR goggles on him and run Madden and tell him he's been accepted. (Some would call this tactic a symptom of the feminization of management.)

Of course, a person in a bubble like that might actually become more radicalized. And if a person discovers he's been had by looking at, say, his friend's Reddit version, he might become angry enough to do something, angrier even than if he had just been told "no", because now he's been made a fool.

Given what we know about how people can end up radicalized in echo chambers with like-minded humans, auto-ensconcing outliers in an echo chamber of fake respondents agreeing with them sounds like a recipe for fomenting a lot of stochastic terrorism.

I independently came up with this idea. but not this way to name it.

also, why would the AIs give only praise? they could also be mean and evil; would that be "hellbanning"?

Besides the obvious issue of operating cost, this seems like it would just further reinforce the undesirable behavior.

Seems like a great way to build extremism.

never heard of that

is that proven? how did they find about that?

is it reddit doing it, or is it other users?

EDIT: ok so that's a fake screenshot, the users says it is... sorry but this is misleading.

he says hypothetical, and then makes a fake screenshot with the NYT on it, that's really disingenuine.

twitter gonna twitter

Wait is this real? So glad I left Reddit, that's really fucked up.

Link goes to a “something went wrong” error page.

That's a riot!

Yeah, it could definitely happen.