Reply to post: this is progress :)

Kill animals and destroy property before hurting humans, Germany tells future self-driving cars

Anonymous Coward
Anonymous Coward

this is progress :)

compared to, say, the 'logic' presented in https://www.theregister.co.uk/2017/07/05/selfdriving_algorithms_that_make_ethical_decisions_based_on_impulse/

or

https://www.theregister.co.uk/2016/11/18/helping_autonomous_vehicles_and_humans_share_the_road/

and the rather worrying simulation offered (linked-to) at http://moralmachine.mit.edu/ which I remember giving up on after less than a half-dozen scenarios ... because it just felt wrong ...

To put it another way, we can't expect a machine to solve a problem, (the trolley problem) that we don't know how to solve, (that's another way to state the Church-[Taski-]Turing thesis, isn't it ?)

We can't expect 'machine learning' to 'just do this' without much more & better real world knowledge than we (humans) could realistically acquire, process and apply within very short time-scales ...

That, I think & hope, is the edge that AI systems should have ...

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2019