@Futilitarian you're right, the feasibility of an 'off switch' depends on how the ai is implemented, and truly no matter how it's implemented, there's probably a path for any true superintelligent general AI to get control of its own off switch eventually.
↧