

Well yeah exactly why I said “the same risk”. ideally it’s going to be in the same systems… and assuming no one is stupid enough (or the laws don’t let them) attach it to the publicly accessible forms of existing AIs It’s not a new additional risk, just the same one. (though those assumptions are largely there own risks.
reminds me a lot of the polar express uncanny valley.
https://i.imgur.com/SW2ch8T.png