Doug Anderson:
> And yet, essentially none of the tools we use obey any of Asimov's
> "laws."
Nature is endlessly inventive and is always finding new and ingenious
ways to kill us. Nonetheless, humans are pretty smart, and we have
learned over the past 100k years or so how to minimize the risk that we
incur from using the tools that we have invented. They can't be made
completely safe, and those who design them do not have Asimov's laws in
mind, but they generally follow an analog of those laws.
> They are capable of harming people, they aren't capable of protecting
> themselves from harm, and they aren't capable of prioritizing human
> protection over their own protection.
Only the first clause is true. Machines are capable of shutting
themselves down, sometimes very quickly, to prevent damage to
themselves, and they are also capable of sacrificing themselves to
protect humans. See <
http://www.youtube.com/watch?v=FquL0GG9RGI> for a
device that can stop a table-saw blade in under 5 milliseconds,
preventing digits from being severed. The designer made the device
prioritize human protection over its own survival; if activated by
contact between the blade and human flesh the device is necessarily
destroyed and must be replaced before the saw can be used again. I know
two fellow woodworkers who accidentally put this device to the test,
and it passed the test in both instances.
> So regardless of what one wants from robots, if you expect them
> (somehow) to be able to satisfy Asimov's laws, there is no precedent
> for that.
They will obey the instructions of their makers or users. Think of
robotic devices on cars that monitor solid lines painted on the road
and warn the driver when the car is too close to the line. My car can
maintain a set difference from the car in front of it and slow itself
when the car in front slows, and resume speed when he resumes speed. If
the car in front slows too rapidly my car sounds an alarm. Some luxury
cars can make autonomous emergency stops when needed. Cars have long
been able detect skid within milliseconds and apply braking to any of
four wheels to prevent skids. Think about it, and you will realize that
robotic devices protect people every day.
Think, too, of CIA drones for an obvious example of robots used to kill
people. One rationale the CIA applies is that these drones save the
lives of blue pilots. While many people question the morality of this
program, no drone has yet questioned the morality of this concept and
either refused to fly on moral grounds or turned on its masters.