[Technology] Self-Driving Uber Car Kills Pedestrian in Arizona.

[Stat Brother]

Arizona officials saw opportunity when Uber and other companies began testing driverless cars a few years ago. Promising to keep oversight light, they invited the companies to test their robotic vehicles on the state’s roads.

Then on Sunday night, an autonomous car operated by Uber — and with an emergency backup driver behind the wheel — struck and killed a woman on a street in Tempe, Ariz. It was believed to be the first pedestrian death associated with self-driving technology. The company quickly suspended testing in Tempe as well as in Pittsburgh, San Francisco and Toronto.

https://www.nytimes.com/2018/03/19/technology/uber-driverless-fatality.html

 

[Brian Parsons]

Stable door and horse spring to mind.

Sent from my SM-G925F using Tapatalk

 

[Goldstone1976]

From what I’ve read, the pedestrian was apparently jay walking and just stepped out in front of the car. I think the key question to be answered is whether a human driver would have avoided her, or hit her less hard. The fact that there was a human in the car too - to act as an emergency backup - would suggest perhaps not.

Whether it’ll be able to be proved that the driverless car performed no worse than one with a driver is open to significant doubt though.

 

[hans kraay fan club]

Arizona officials saw opportunity when Uber and other companies began testing driverless cars a few years ago. Promising to keep oversight light, they invited the companies to test their robotic vehicles on the state’s roads.

Then on Sunday night, an autonomous car operated by Uber — and with an emergency backup driver behind the wheel — struck and killed a woman on a street in Tempe, Ariz. It was believed to be the first pedestrian death associated with self-driving technology. The company quickly suspended testing in Tempe as well as in Pittsburgh, San Francisco and Toronto.

https://www.nytimes.com/2018/03/19/technology/uber-driverless-fatality.html

An absolute tragedy, on an individual level, of course.

But how many miles of successful testing have they accomplished. How many pedestrians have been killed in Arizona over the same timescale, by cars driven by humans? A lot more than one, I'd imagine.

Nobody should be writing off the technology or the concept, off the back of this. It will, in time, save a lot of lives.

 

[Uter]

It will, in time, save a lot of lives.

And money. From memory, I believe the NHS spends around £8billion a year dealing with RTA casualties.

 

[Goldstone1976]

Assuming that basics of getting a driverless car to work properly are sorted at some point, I think there’s a real philosophical challenge around all AI-controlled automation that impacts on human safety.

How do you teach it intuition &/or ethics? Say that a car had to make a decision between ramming another car head on, driving onto the pavement where it’ll hit a mother pushing a pram, driving straight into a tree on the other side of the road, or trying to weave through the oncoming traffic with a 1% chance of pulling it off. A human has intuition and thousands of generations of societal pressure influencing their decision-making process. How does that get embedded into an AI?

It’s (relatively) easy to programme it to do its best to avoid all accidents; much harder to get it to choose the least bad option when an accident is inevitable, especially if what constitutes “least bad” is debatable...

 

[strings]

Assuming that basics of getting a driverless car to work properly are sorted at some point, I think there’s a real philosophical challenge around all AI-controlled automation that impacts on human safety.

How do you teach it intuition &/or ethics? Say that a car had to make a decision between ramming another car head on, driving onto the pavement where it’ll hit a mother pushing a peak, driving straight into a tree on the other side of the road, or trying to weave through the oncoming traffic with a 1% chance of pulling it off. A human has intuition and thousands of generations of societal pressure influencing their decision-making process. How does that get embedded into an AI?

It’s (relatively) easy to programme it to do its best to avoid all accidents; much harder to get it to choose the least bad option when an accident is inevitable, especially if what constitutes “least bad” is debatable...

This, for me, is everything about AI. At some point, it will have to make life-or-death decision where it finds itself in a position where crashes are unavoidable. This is a real ethical minefield.

 

[BLOCK F]

I heard on LBC that the car was speeding!
May be I am mising something, but I can't see the point in this self drive car lark.
What is the real purpose?

 

[Jim D]

I heard on LBC that the car was speeding!
May be I am mising something, but I can't see the point in this self drive car lark.
What is the real purpose?

More profits for Uber, cars operating 24/365, no driver strikes.....

 

[el punal]

I heard on LBC that the car was speeding!
May be I am mising something, but I can't see the point in this self drive car lark.
What is the real purpose?

On a flippant note, if it replaced taxi drivers at least you won't get the "You'll never guess who I had in the back of the cab?" chat fest.

 

[Silverhatch]

On a flippant note, if it replaced taxi drivers at least you won't get the "You'll never guess who I had in the back of the cab?" chat fest.

You probably would - but in a Stephen Hawking type voice

 

[Stat Brother]

You probably would - but in a Stephen Hawking type voice

 

[Gwylan]

I heard on LBC that the car was speeding!
May be I am mising something, but I can't see the point in this self drive car lark.
What is the real purpose?

There are going to be huge benefits. Fewer accidents will mean less congestion and lower NHS costs for one thing. No labour costs will mean rides are going to be cheaper but, most of all, it will mean a more efficient use of road space.

We won't see autonomous vehicles as the norm for some years yet but I fully expect to see the end of people driving their own cars in my lifte-time (and I'm in my 60s)

 

[lost in london]

This, for me, is everything about AI. At some point, it will have to make life-or-death decision where it finds itself in a position where crashes are unavoidable. This is a real ethical minefield.

I would back the AI on balance - have you not seen the variety of reactions us humans have? Something as stupid as a jar dropping off a worktop - some will try and catch it (me), others try and stop it with their foot (also me - always a weird reaction that one), others scream, jump away and shut their eyes (my wife). At least the AI would be consistent in how it reacts and perhaps just try and stop as quickly as possible (almost certainly hitting the brakes harder and quicker than any human).

 

[strings]

I would back the AI on balance - have you not seen the variety of reactions us humans have? Something as stupid as a jar dropping off a worktop - some will try and catch it (me), others try and stop it with their foot (also me - always a weird reaction that one), others scream, jump away and shut their eyes (my wife). At least the AI would be consistent in how it reacts and perhaps just try and stop as quickly as possible (almost certainly hitting the brakes harder and quicker than any human).

I agree (and I do the strange 'cushion a falling object with my foot, so that it doesn't hit the floor too hard' thing too!). AI will be more reliable, more consistent. It is just a strange thought that in any given set of circumstance, a programmer (or a committee perhaps) has pre-determined the outcome of the scenario.

 

[lost in london]

I heard on LBC that the car was speeding!
May be I am mising something, but I can't see the point in this self drive car lark.
What is the real purpose?

Drive on the M25 in the morning - three or four lanes of cars, all pretty full, all driving at the same(ish) speed. Infinitely safer if cars were constantly monitoring distance from cars in front and behind. With humans it takes just one person to hit the brakes a bit hard and queues develop:

https://www.youtube.com/watch?v=Suugn-p5C1M

But research suggests you don't need many autonomous vehicles (one will help) to overcome this:

https://www.technologyreview.com/s/...car-has-a-huge-impact-on-alleviating-traffic/

An interesting side benefit is greater fuel efficiency / less pollution.

Where I think it will struggle is driving round our weird little windy towns and villages, with cars parked on curbs, vans pulling over to make deliveries, people flashing lights at another car to come through a tight spot etc. God knows how they program the car to deal with that.

 

[lost in london]

I agree (and I do the strange 'cushion a falling object with my foot, so that it doesn't hit the floor too hard' thing too!). AI will be more reliable, more consistent. It is just a strange thought that in any given set of circumstance, a programmer (or a committee perhaps) has pre-determined the outcome of the scenario.

The other big downside will be the general falling of driving standards - the more technology we have in cars, the less we think we need to pay attention. The transition period to the inevitable autonomous driving might be a messy one.

It's an interesting thought though that my two young kids (under 7) might feasibly never need to learn how to drive.

 

[Bold Seagull]

Apparently AI self driven cars all use indicators when required.

This is an incredible leap forward given how many human beings cannot master that skill.

 

[Wellesley]

Still safer than being driven by their sex-offending, terrorist drivers. What?

 

[Stat Brother]

Apparently AI self driven cars all use indicators when required.

This is an incredible leap forward given how many human beings cannot master that skill.

At least it's nice to know which technology isn't being driven (pun intended) by Germany.