Killer robots

Question 1

What are the different degrees of machine autonomy in weapons

Answer 1

1. human-in-the-loop
2. human-on-the-loop
3. human-out-of-the-loop

Question 2

Human-in-the-loop

Answer 2

machine recommends targets, but a human soldier must decide whether to engage the target

Question 3

Human-on-the-loop

Answer 3

machine selects and engage targets with a human monitoring and able to intervene

Question 4

Human-out-of-the-loop

Answer 4

machine selects and engages targets without human oversight

Question 5

ICRC definition of lethal autonomous weapons

Answer 5

any system that automates targeting and engaging a weapon

Question 6

what does the ‘just war’ theory assume?

Answer 6

war is a great evil and impermissible by default.
specific exceptions are necessary to authorise the use of force

Question 7

Main dimensions of ‘just war’ theory

Answer 7

1.lus ad bellum
2.lus in bello
3. lus post bellum

Question 8

Lus ad bellum

Answer 8

covers conditions under which it is permissible to go to war

Question 9

Lus in bello

Answer 9

covers permissible conduct in war

Question 10

lus post bellum

Answer 10

covers permissible conduct in engine war and post-war relations

Question 11

Lus ad bellum: conditions for authorising armed force

Answer 11

just cause: was may only be initiated for self-defence and the defence of others from aggression
last resort: non-violent options must be exhausted before force can be justified
probability of success: the times of war must have a high likelihood of achievement
war must be publicly initiated by a state

Question 12

lus in bello: conditions for conduct in war

Answer 12

distinction: acts of war should target enemy combatants and not non-combatants
military necessity: attacks must advance legitimate military objectives
proportionality: any collateral damage must be proportional to legitimate military objectives
inherently evil tactics may not be used

Question 13

Objections to lethal uninhabited aerial vehicles (drones) Strawser ‘Moral predators’ (2010)

Answer 13

invites indiscriminate and disproportionate attackts
create unjust asymmetry in combat
create cognitive dissonance for operators, lowering their responsibility threshold
slippery slope to autonomous weapons

Question 14

Military conduct in Operation Iraqi Freedom (Arkin 2009)

Answer 14

small amount of soldiers reporting mistreatment of noncombatants
Arkin: it is conceptually and mathematically possible to develop a class of robots that not only conform to international law but outperform human soldiers in their ethical capacity

Question 15

Asaro on autonomous weapons (2020)

Answer 15

building and using autonomous weapons is morally wrong on both deontological and consequentialists grounds

Question 16

Deontological arguments

Answer 16

responsibility - performance of moral obligations delegated to other moral agents who take responsibility. Wrong to delegate to LAWS because they are not moral agents
dignity - decisions for lethal force must recognise a human being as a human, reach a rational conclusion that killing is justified in particular situations

Question 17

Consequentialist arguments

Answer 17

LAWS might accidentally misidentify targets or intentionally designed with these purposes
LAWS don’t require sophisticated technology and can be easily acquired by non state actors
LAWS may accidentally initiate or escalate conflicts; the interaction between adversarial systems is unpredictable and may be catastrophic
LAWS may lower the threshold for initiating or escalating military conflicts
LAWS may engage in unattributable attacks, leading to widespread assassinations or random attacks to destabilise societies
LAWS are susceptible to spoofing and hacking; tricked into seeing friends as enemies and stop signs as go signs