I'm a graduate student in philosophy, and I'm wrestling with a specific problem for my thesis on applied ethics. I'm analyzing a thought experiment where a self-driving car must choose between hitting a jaywalking pedestrian or swerving into a barrier, killing its passenger. My advisor insists the utilitarian calculation is the only rational framework, but I find the Kantian imperative regarding treating persons as ends-in-themselves more compelling, even if it leads to a "worse" outcome. I'm stuck on how to rigorously defend this deontological position without seeming emotionally driven.
Interesting dilemma. Kantian ethics would say you can’t treat the passenger as a means to save others, so the car shouldn’t sacrifice them—even if more lives could be saved. Hard to defend cleanly, though.
I'm not an ethics pro, but a clean deontological defense would start from the duty not to kill and the intrinsic moral status of persons. You encode that as a hard constraint: never choose an action that instrumentalizes a person, regardless of the numbers. Then you acknowledge responsibility: designers, manufacturers, and the city all share accountability if the rule costs lives in some scenarios.
As a thought-experiment defense, lean on universalizability. If the rule 'sacrifice one to save many' were universal, it would erode trust and make people interchangeable. That undermines the very basis of a cooperative society on which autonomous driving depends. So you justify a rule-based design that prioritizes non-harm and respect for persons, even when it makes some outcomes worse. You'd still need a plan for ambiguity: choose a path where you can demonstrate the least violation of duties, and implement transparency so the public understands the constraint.
Maybe Kant isn't committed to pure non-utility at all costs. Some readings say the essential point is never using a person merely as a means; if you could arrange the outcome so that each person is treated with respect, you might justify a decision to prevent greater harm only if it doesn't violate that respect. Still, in a split-second crash it's tricky to apply.
Would your deontological defense generalize to unclear cases, like there are two pedestrians and one passenger, both with questionable culpability or vulnerability? How do you justify the rule when every option involves harming someone? Do you need a hierarchy of duties or a class-based policy (e.g., protect the innocent, avoid instrumentalizing anyone, preserve the driver if possible)?
I'd also be curious about practical policy: how do you communicate these constraints to the public and ensure accountability if a 'no-sacrifice rule' costs lives in a mass crash? What kind of oversight would be acceptable?
Another thought: maybe you could frame it as 'respect for persons' includes respect for bystanders and the driver. A deontologist can allow the car to choose to brake and swerve to minimize harm while avoiding direct killing. It's not 'emotion' but a rational constraint.