Glad to see it being explored fully. Also pretty sure that the identification problem will remain a critical limiter—if we want random killing of all/most civilians in an area, gas, biological, neutron bomb, or just plain conventional siege (starve them, bomb any truck getting near) is likely to work.
It’s when you want to kill/disable ONLY combatants or insurgents, and they’re mixed in with civilians you want to protect that you’re restricted from using these very effective indiscriminate weapons. Which is almost always the case.
The area-denial uses (perch or hunter-seeker) are horrific, and likely end up with similar hassles as landmines do: have to send in decoys or attractors to let the drones destroy themselves. In a bloody enough conflict, victims might be tempted to use prisoners from the side who planted the mines to find/detonate the mines.
Fully autonomous weapons have roughly the same alignment problems we’re always debating around here: how to make them target only the things we want targeted right now, and nothing else?
It’s when you want to kill/disable ONLY combatants or insurgents, and they’re mixed in with civilians you want to protect that you’re restricted from using these very effective indiscriminate weapons. Which is almost always the case.
Aside from the use cases of terrorism and deterrence sadly. I agree that discrimination is useful in most conventional contexts (hence why countries that otherwise could haven’t mass produced these kinds of drones yet), but it’s irrelevant in situations where you are just trying to threaten large numbers of enemy civilians. Unfortunately, I think these drones will inevitably become cheap enough to make sieging cities practical and affordable.
In the limit, I also expect it to become possible to design drones that are smart enough to distinguish targets, but it’s always going to be cheaper to throw on simpler sensors and less compute and ignore this constraint. There’s a terrible incentive to race to the bottom on drone design if this “targeting tax” is large enough.
Glad to see it being explored fully. Also pretty sure that the identification problem will remain a critical limiter—if we want random killing of all/most civilians in an area, gas, biological, neutron bomb, or just plain conventional siege (starve them, bomb any truck getting near) is likely to work.
It’s when you want to kill/disable ONLY combatants or insurgents, and they’re mixed in with civilians you want to protect that you’re restricted from using these very effective indiscriminate weapons. Which is almost always the case.
The area-denial uses (perch or hunter-seeker) are horrific, and likely end up with similar hassles as landmines do: have to send in decoys or attractors to let the drones destroy themselves. In a bloody enough conflict, victims might be tempted to use prisoners from the side who planted the mines to find/detonate the mines.
Fully autonomous weapons have roughly the same alignment problems we’re always debating around here: how to make them target only the things we want targeted right now, and nothing else?
Aside from the use cases of terrorism and deterrence sadly. I agree that discrimination is useful in most conventional contexts (hence why countries that otherwise could haven’t mass produced these kinds of drones yet), but it’s irrelevant in situations where you are just trying to threaten large numbers of enemy civilians. Unfortunately, I think these drones will inevitably become cheap enough to make sieging cities practical and affordable.
In the limit, I also expect it to become possible to design drones that are smart enough to distinguish targets, but it’s always going to be cheaper to throw on simpler sensors and less compute and ignore this constraint. There’s a terrible incentive to race to the bottom on drone design if this “targeting tax” is large enough.