ai ethics

June 2026

can we teach ai to be ethical? what does it mean for ai to be ethical?

ethical dilemma goes like this: programming a self-driving car. has to make a choice between hitting two cyclists, one who is wearing a helmet, one who isn't. which should we program the car to hit? the one who is abiding by the rules or the one who is less likely to survive?

in my opinion, there are two veins of thought:

a) we act as a utilitarian, minimising the expected loss of life

b) we act as the law, taking it into our own hands to punish those who don't abide by the rules.

at the moment, i am not sure where i stand, potentially a mix of both.

the question can be extended further: if a), then what happens if one cyclist is healthy/disabled/more or less likely to survive? also, what if one of them is a murderer/rapist/criminal vs someone just not wearing a helmet?

lots to digest.