characteristics of automated human kill - deferred executive decision (the "deploy order") - multiplier on the executive decision: one decision-->multiple kill outcomes - non-specificity of the target - a machine has no "heart & soul" in its programming - machines can be constructed & deployed with an economy of scale
reasons for constraints on methods of war (E.g., Geneva Protocol & Hague Conventions) - limit excessive & needless human suffering - limitation of indiscriminate target kills
hazards - automation of killing can provide a cost-effective expediency free of moral hazard to the operator - expediency removes incentive to find alternative, non-lethal yet more costly solutions - an operator may not perceive complicity of his role in the kill ("cleansing of hands") - whereas the executive order to kill multiple, non-specific human targets has in past been issued by a professional commander--the risk is that ease-of-implementation and training will lead to trained operators taking the place of experienced commanders with "heart & soul" experience of the battlefield. - because the autonomous machine has no "heart & soul" it will not deviate from programming if confronted with a new situation and may not exercise restraint when such course would be prudent. Example--software may not recognize a house used for child-care when field commander issues order to the robot bulldozer team "shortest path" - because the autonomous machine has no "heart & soul" the machine will not recognize when a bad, or imprudent order has been given, or an order given with criminal intent. perhaps a "dual key" system that requires two officers to agree before deployment would mitigate this aspect. - because the autonomous machine has no "heart & soul" a robot army in concept could be configured to assault the homeland without knowing why it shouldn't.
62° Partly Cloudy
Pros and Cons of Robotic Warfare
some thoughts:
characteristics of automated human kill
- deferred executive decision (the "deploy order")
- multiplier on the executive decision: one decision-->multiple kill outcomes
- non-specificity of the target
- a machine has no "heart & soul" in its programming
- machines can be constructed & deployed with an economy of scale
reasons for constraints on methods of war (E.g., Geneva Protocol & Hague Conventions)
- limit excessive & needless human suffering
- limitation of indiscriminate target kills
hazards
- automation of killing can provide a cost-effective expediency free of moral hazard to the operator
- expediency removes incentive to find alternative, non-lethal yet more costly solutions
- an operator may not perceive complicity of his role in the kill ("cleansing of hands")
- whereas the executive order to kill multiple, non-specific human targets has in past been issued by a professional commander--the risk is that ease-of-implementation and training will lead to trained operators taking the place of experienced commanders with "heart & soul" experience of the battlefield.
- because the autonomous machine has no "heart & soul" it will not deviate from programming if confronted with a new situation and may not exercise restraint when such course would be prudent. Example--software may not recognize a house used for child-care when field commander issues order to the robot bulldozer team "shortest path"
- because the autonomous machine has no "heart & soul" the machine will not recognize when a bad, or imprudent order has been given, or an order given with criminal intent. perhaps a "dual key" system that requires two officers to agree before deployment would mitigate this aspect.
- because the autonomous machine has no "heart & soul" a robot army in concept could be configured to assault the homeland without knowing why it shouldn't.
September 2, 2009 at 9:49 a.m. ( permalink | suggest removal )