“How can you control something that can actually reprogram itself?”
I’ve been reading up on drone technology recently and it’s pretty obvious that we’ll see the first Terminator in our life time – perhaps not in human form as he’ll be steadier and faster on 4 legs than 2, perhaps with wheels too or who knows what. Many would say the current unmanned drones are terminators – they kill regularly but they’ve not yet been given autonomy to decide themselves when to use lethal force. Autonomy is the controversial next step.
In this very interesting book ‘Wired for War‘, PW Singer explains the different levels of autonomy:
Direct human operation: having some guy behind a computer control all of the plane’s operations from the ground.
Human-assisted: when the pilot on the ground takes off and lands the plane, but can let the plane fly itself while in the air.
Human Delegation: the pilot has to instruct the plane to take off and land and give it way-points to fly to.
Human-supervised: the operator is really no longer a pilot, but monitors what information it sends back.
Mixed-initiative: the human might give the robotic plane a mission to accomplish, but doesn’t need to oversee it. The plane is given a mission file to complete or even loosely told to be “curious” and only report back when it finds something interesting.
Fully-autonomous: the machine decides on its own what to report and where to go.
Adaptive: a machine is adaptive when it can learn, update or change what it should search out or even evolving to gather information in new ways.
Update 14 July 2013
It seems the Terminator’s name is “Atlas” reports the BBC: