It's a terrible idea in more ways than one. The biggest being the fact that AI = software = will be hacked. And that excludes scenarios where country elites wi

It's a terrible idea in more ways than one. The biggest being the fact that AI = software = will be hacked. And that excludes scenarios where country elites will control it from behind the scenes with algorithm changes that favor them, while pretending to improve it. Maybe all of these issues will be solved in 100 years. But I think it's far too premature to actually hope for this happening soon. It's going to end in disaster if it somehow does happen, because people are tricked into thinking it's a great idea (just like Pentagon officials and politicians are not being tricked by war contractors that autonomous weapons is a good idea with the AI capabilities and software security we have today). Because of this trickery we're going to see increasingly more bad decisions by governments who allow AI to take their place. Such as this recent one: https://www.techdirt.com/articles/20180409/09125639594/uk-police-use-zipcode-profiles-garden-size-first-names-ai-based-custody-decision-system.shtml https://www.techdirt.com/articles/20180409/09125639594/uk-po...