Thank you for this contribution and for all the work that this article required.
I will try to go in order through the arguments that I find most interesting to further discuss.
A) Responsibility:
What is the evidence that human societies are capable of holding responsible generals and technology producers for death caused by a certain technology?
If I think about the poor record the International Criminal Court has of bringing war criminals to justice, and the fact that the use of cluster bombs in Laos or Agent Orange in Vietnam did not lead to major trials, I am skeptical on whether someone would be hold accountable for crimes committed by LAWs.
2. What evidence do we have that international lawmaking follows suit when a lethal technology is developed as the writer assumes it will happen?
From Wikipedia on Arms Control: “According to a 2020 study in the American Political Science Review, arms control is rare because successful arms control agreements involve a difficult trade-off between transparency and security. For arms control agreements to be effective, there needs to be a way to thoroughly verify that a state is following the agreement, such as through intrusive inspections. However, states are often reluctant to submit to such inspections when they have reasons to fear that the inspectors will use the inspections to gather information about the capabilities of the state, which could be used in a future conflict.”
B) Moral Hazard:
The article does a good work in looking at the evidence available for ratio military/civilian casualties in war.
However, in order for the comparison to make more sense I would argue that the different examples should be weighted according to the number of victims.
Intuitively to me, the case for LAWs increasing the chance of overseas conflicts such as the Iraq invasion is a very relevant one, because of the magnitude of civilian deaths.
From what the text says I do not see why the conclusion is that banning LAWs would have a neutral effect on the likelihood of overseas wars, given that the texts admits that it is an actual concern.
I think the considerations about counterinsurgencies operations being positive for the population is at the very least biased towards favoring Western intervention.
One could say that if a state like Iraq was not destabilized in the first place, it would not be ground for the expansion of groups such as ISIL.
Twenty years and 38,000 civilian deaths after the beginning of the Afghanistan War, the Talibans still control half of the country.
C) Domestic conflict
I think this section makes lots of assumptions and this stresses the high level of uncertainty we have about this topic, I will just quote some of the passages that are open to debate:
“Their new military capability likely won’t change much, as rebels could obtain their own AI and rebels have usually taken asymmetric approaches with political and tactical exploits to circumvent their dearth of heavy organized combat power.”
“If autocrats make errant over-use of LAWs to protect their regimes, as many states have done with regular mechanized forces, they could inadvertently weaken themselves. Otherwise, smart states will seek to avoid being forced to deploy LAWs against their own people.”
“Future technologies may increase popular power even further.” → the rise of social media and conspirationism does not necessarily constitute an increase in popular power.
D) Effects of campaigning against killer robots
The considerations about China and the world order in this section seem simplistic and rely on many assumptions.
If I think about the poor record the International Criminal Court has of bringing war criminals to justice, and the fact that the use of cluster bombs in Laos or Agent Orange in Vietnam did not lead to major trials, I am skeptical on whether someone would be hold accountable for crimes committed by LAWs.
But the issue here is whether responsibility and accountability is handled worse with LAWs as compared with normal killing. You need a reason to be more skeptical for crimes committed by LAWs than you are for crimes not committed by LAWs. That there is so little accountability for crimes committed without LAWs even suggests that we have nothing to lose.
What evidence do we have that international lawmaking follows suit when a lethal technology is developed as the writer assumes it will happen?
I don’t think I make such an assumption? Please remind me (it’s been a while since I wrote the essay), you may be mistaking a part where I assume that countries will figure out safety and accountability for their own purposes. They will figure out how to hold people accountable for bad robot weapons just as they hold people accountable for bad equipment and bad human soldiers, for their own purposes without reference to international laws.
However, in order for the comparison to make more sense I would argue that the different examples should be weighted according to the number of victims.
I would agree if we had a greater sample of large wars, otherwise the figure gets dominated by the Iran-Iraq War, which is doubly worrying because of the wide range of estimates for that conflict. You could exclude it and do a weighted average of the other wars. Either way, seems like civilians are still just a significant minority of victims on average.
Intuitively to me, the case for LAWs increasing the chance of overseas conflicts such as the Iraq invasion is a very relevant one, because of the magnitude of civilian deaths.
Yes, this would be similar to what I say about the 1991 Gulf War—the conventional war was relatively small but had large indirect costs mostly at civilians. Then, “One issue with this line of reasoning is that it must also be applied to alternative practices besides warfare...” For Iraq in particular, while the 2003 invasion certainly did destabilize it, I also think it’s a mistake to think that things would have been decent otherwise (imagine Iraq turning out like Syria in the Arab Spring; Saddam had already committed democide once, he could have done it again if Iraqis acted on their grievances with his regime).
From what the text says I do not see why the conclusion is that banning LAWs would have a neutral effect on the likelihood of overseas wars, given that the texts admits that it is an actual concern.
My ‘conclusion’ paragraph states it accurately with the clarification of ‘conventional conflicts’ versus ‘overseas counterinsurgency and counterterrorism’
I think the considerations about counterinsurgencies operations being positive for the population is at the very least biased towards favoring Western intervention.
Well, the critic of AI weapons needs to show that such interventions are negative for the population. My position in this essay was that it’s unclear whether they are good or bad. Yes, I didn’t give comprehensive arguments in this essay. But since then I’ve written about these wars in my policy platform where you can see me seriously argue my views, and there I take a more positive stance (my views have shifted a bit in the last year or so).
The considerations about China and the world order in this section seem simplistic and rely on many assumptions.
Once more, I got you covered! See my more recent essay here about the pros and cons (predominately cons) of Chinese international power. (Yes it’s high time that I rewrote and updated this article)
Thank you for this contribution and for all the work that this article required.
I will try to go in order through the arguments that I find most interesting to further discuss.
A) Responsibility:
What is the evidence that human societies are capable of holding responsible generals and technology producers for death caused by a certain technology?
If I think about the poor record the International Criminal Court has of bringing war criminals to justice, and the fact that the use of cluster bombs in Laos or Agent Orange in Vietnam did not lead to major trials, I am skeptical on whether someone would be hold accountable for crimes committed by LAWs.
2. What evidence do we have that international lawmaking follows suit when a lethal technology is developed as the writer assumes it will happen?
From Wikipedia on Arms Control: “According to a 2020 study in the American Political Science Review, arms control is rare because successful arms control agreements involve a difficult trade-off between transparency and security. For arms control agreements to be effective, there needs to be a way to thoroughly verify that a state is following the agreement, such as through intrusive inspections. However, states are often reluctant to submit to such inspections when they have reasons to fear that the inspectors will use the inspections to gather information about the capabilities of the state, which could be used in a future conflict.”
B) Moral Hazard:
The article does a good work in looking at the evidence available for ratio military/civilian casualties in war.
However, in order for the comparison to make more sense I would argue that the different examples should be weighted according to the number of victims.
Intuitively to me, the case for LAWs increasing the chance of overseas conflicts such as the Iraq invasion is a very relevant one, because of the magnitude of civilian deaths.
From what the text says I do not see why the conclusion is that banning LAWs would have a neutral effect on the likelihood of overseas wars, given that the texts admits that it is an actual concern.
I think the considerations about counterinsurgencies operations being positive for the population is at the very least biased towards favoring Western intervention.
One could say that if a state like Iraq was not destabilized in the first place, it would not be ground for the expansion of groups such as ISIL.
Twenty years and 38,000 civilian deaths after the beginning of the Afghanistan War, the Talibans still control half of the country.
C) Domestic conflict
I think this section makes lots of assumptions and this stresses the high level of uncertainty we have about this topic, I will just quote some of the passages that are open to debate:
“Their new military capability likely won’t change much, as rebels could obtain their own AI and rebels have usually taken asymmetric approaches with political and tactical exploits to circumvent their dearth of heavy organized combat power.”
“If autocrats make errant over-use of LAWs to protect their regimes, as many states have done with regular mechanized forces, they could inadvertently weaken themselves. Otherwise, smart states will seek to avoid being forced to deploy LAWs against their own people.”
“Future technologies may increase popular power even further.” → the rise of social media and conspirationism does not necessarily constitute an increase in popular power.
D) Effects of campaigning against killer robots
The considerations about China and the world order in this section seem simplistic and rely on many assumptions.
Hi Tommaso,
But the issue here is whether responsibility and accountability is handled worse with LAWs as compared with normal killing. You need a reason to be more skeptical for crimes committed by LAWs than you are for crimes not committed by LAWs. That there is so little accountability for crimes committed without LAWs even suggests that we have nothing to lose.
I don’t think I make such an assumption? Please remind me (it’s been a while since I wrote the essay), you may be mistaking a part where I assume that countries will figure out safety and accountability for their own purposes. They will figure out how to hold people accountable for bad robot weapons just as they hold people accountable for bad equipment and bad human soldiers, for their own purposes without reference to international laws.
I would agree if we had a greater sample of large wars, otherwise the figure gets dominated by the Iran-Iraq War, which is doubly worrying because of the wide range of estimates for that conflict. You could exclude it and do a weighted average of the other wars. Either way, seems like civilians are still just a significant minority of victims on average.
Yes, this would be similar to what I say about the 1991 Gulf War—the conventional war was relatively small but had large indirect costs mostly at civilians. Then, “One issue with this line of reasoning is that it must also be applied to alternative practices besides warfare...” For Iraq in particular, while the 2003 invasion certainly did destabilize it, I also think it’s a mistake to think that things would have been decent otherwise (imagine Iraq turning out like Syria in the Arab Spring; Saddam had already committed democide once, he could have done it again if Iraqis acted on their grievances with his regime).
My ‘conclusion’ paragraph states it accurately with the clarification of ‘conventional conflicts’ versus ‘overseas counterinsurgency and counterterrorism’
Well, the critic of AI weapons needs to show that such interventions are negative for the population. My position in this essay was that it’s unclear whether they are good or bad. Yes, I didn’t give comprehensive arguments in this essay. But since then I’ve written about these wars in my policy platform where you can see me seriously argue my views, and there I take a more positive stance (my views have shifted a bit in the last year or so).
Once more, I got you covered! See my more recent essay here about the pros and cons (predominately cons) of Chinese international power. (Yes it’s high time that I rewrote and updated this article)