[Tyler]

Driverless Cars Will Face Moral Dilemmas
23 June 2016, 12:00 am

                    A self-driving car carrying a family of four on a rural two-lane highway spots a bouncing ball ahead. As the vehicle approaches a child runs out to retrieve the ball.

Scientific American: TechnologyLink

Source: AI in the News

To visit any links mentioned please view the original article, the link is at the top of this post.

[ivan.moony]

If the bot has to kill anyone, someone is not doing his job right. I guess putting cars on dedicated 10 meters altitude tracks, where no walkers can approach, is too expensive. It is cheaper to occasionally kill someone.

[Art]

Well any decent auto program should immediately slow the car if an object appears at, near or in a possible collision path with the auto. Perhaps it should go into a "creep" mode until sufficient time has elapsed or a person (child, animal) appears in the line of vision or projected direction.
Yeah...it will cost a few more dollars of programming time to implement but how can one place a price on a life?

What happens in thick, dense fog, smog or a pounding rainstorm or smoke or very overcast days when the GPS and car mounted cameras cannot "see" well? Does the car stop? If there are no operator overrides, then you're just stuck sitting in a four-wheeled box, waiting for conditions to clear or improve. Cool!

[Don Patrick]

It's not a moral dilemma, it's called an unavoidable accident. I don't think people carefully consider the ethics of it all before slamming the brakes.
You see a kid, you brake to the degree that it doesn't crash the car, you honk the horn, and hope for the best. Computers have a 1 second head start on human response times to begin with, and they probably brake for balls too.