The problem with AI

People talk about the problem of artificial intelligence as the risk of AI entities thinking for themselves and acting against their human masters. That may be a problem, but I see a much more obvious one.

Every despot stays in power only so long as he can command violence or the threat of violence. Pushed too far there is always the risk that even the Imperial Guard will turn against the despot. History teaches us that many tyrants are poor judges of when the tipping point is near. But poor judges or not, tyrants are aware that there is a tipping point. Many will play with pushing things to the limit. But they are nonetheless aware the tide can turn against them.

And that knowledge curtails their worst excesses. Or if they push on regardless as some tyrants have done, then history teaches that that can bring on a revolution. Either way, behaviour is modified.

But take away the conscience from the Imperial Guard – replace it with robots – and there will be no end to tyranny.