[ExI] Self-Driving Cars Must Make Ethical Decisions

[Tara Maya]

My prediction is that the biggest danger of self-driving cars will be if they are susceptible to being hacked by malicious humans. Harm will be done on purpose, by trolls or terrorists, far exceeding what will happen by unavoidable accidents. 

Tara Maya
Blog  |  Twitter  |  Facebook  |  Amazon  |  Goodreads



> On Jul 29, 2015, at 8:12 AM, Chris Hibbert <hibbert at mydruthers.com> wrote:
> 
>> Researchers are trying to program self-driving cars to make
>> split-second decisions that raise real ethical questions.
> 
> This seems like a red herring to me. The ethical questions that philosophers (and members of this list) like to debate are all borderline questions, with strong proponents on many sides. If there isn't a single obvious answer to educated discussants sitting quietly in their armchairs, then why is it crucial that automated driving software be able to make split-second distinctions without warning or any background on the potential gains and losses?
> 
> The real issue that needs to be resolved before deciding that it's okay to put AIs in charge of high-speed vehicles is whether they're better at humans at preventing accidents in the incidents that happen every day in every city. And if they're just barely better at that, then any difference of opinion on the subtle ethical trade-offs where just avoiding the accident isn't an option will be very much in the noise.
> 
> Chris
> --
> Rationality is about drawing correct inferences from limited,
>     confusing, contradictory, or maliciously doctored facts.
>  -- Scott Alexander
> 
> Chris Hibbert
> hibbert at mydruthers.com
> http://mydruthers.com
> 
> _______________________________________________
> extropy-chat mailing list
> extropy-chat at lists.extropy.org
> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat