Self driving cars & The Moral Machine

[classic33]

Pretty irrelevant really. Assuming that self driving cars reduce the chances of being involved in an accident to around the same as being struck by lightening the odd anomaly won't really matter.

Never been struck by lightning, cars at present(with driver control), over a dozen times. The last time on a crossing(third time).

Not very good odds. "Lightning Roy" only got hit seven times.

The MIT research is one competitor from the driverless car challenges, publishing their work on a what has become a real challenge for all involved.

 

[NorthernDave]

That is pretty much a given.

But restating the sort of question on the MIT site:

Given a complete brake failure, and only two choices - straight ahead or swerve into the opposite carriage way - what should the car be programmed to do: Kill the 5 people crossing the road in front of it or the cyclist in the opposite carriage way? Does the age, gender, size or criminality of the potential victims influence you choice? If the people are crossing on a Red man and the cyclist is correctly stopped at the crossing is the answer different?

In all probability the car would have detected an issue with the brakes before it became safety critical and have parked itself, requesting assistance. You can already buy mass-market cars that can do this sort of thing.

 

[PK99]

In all probability the car would have detected an issue with the brakes before it became safety critical and have parked itself, requesting assistance. You can already buy mass-market cars that can do this sort of thing.

True. But the moral/philosophical question arises on the occasion it does not. Plus, the MIT questions do raise some interesting dilemmas relating to perceived values of different lives.

 

[swee'pea99]

perceived values of different lives.

Kind of akin to my daughter's lament about equal Brexit votes, regardless of age. As she put it: "People should have had weighted votes to reflect how long they'll have to live with the outcome."

 

[ufkacbln]

Kind of akin to my daughter's lament about equal Brexit votes, regardless of age. As she put it: "People should have had weighted votes to reflect how long they'll have to live with the outcome."

I can just see in a few years time, the "Brenter" Party leading a successful return to Europe after the Brexiteers either pass away or see just how little they actually gained

 

[classic33]

Why does the moral question arise for self driving cars? In situations where an element of decision may have arisen, do people currently get asked about how they decided their lesser of two evils?

Give the car another scenario. It rounds a corner, there's a tanker parked up on one side of the road, a fairly solid wall on the other. There's two cyclists who have moved out, that's why the tanker is stationary, to pass the tanker.

What does the car do? Save its occupants by hitting the wall, or hit the cyclists.

Comes a bit closer to what they're actually trying to solve.

We don't have many drops measured in 1,000's of feet at the side of the road in this country, so the wall has been used instead.

 

[Yellow Saddle]

What is the issue to be solved here? We don't appear to do the moral thing for human drivers, so why invent the issue for machines?

Because it will be written down for everyone to see. That will probably be the first time the exact reasoning for an inconvenient choice will be there for everyone to see and we are more uncomfortable with that than a vague BS rationale.

 

[Yellow Saddle]

You think that it will be written down that faced with x children one side of the road v y OAPs the other and the car will inevitably kill one or the other, then the choice will be x?

Of course, not, it will be in political speak.

 

[classic33]

What is the issue to be solved here? We don't appear to do the moral thing for human drivers, so why invent the issue for machines?

It's something that's happened, that those making the cars never even considered.

And at present there's still some driver control available. Driver decided on going over the edge(hitting the wall), rather than hit the tanker and cyclists.

Google won't publish their work, nor will Tesla, on this bit of research. MIT have, but have no plans to produce self driving vehicles.

 

[gbb]

[QUOTE 4754879, member: 9609"]I wonder how they are working now, would they swerve off the road to avoid a collision of any kind, even an oncoming vehicle on the wrong side of the road, or would they solely concentrate on coming to a halt within their own lane before any collision. i sort of suspect the later as I would like to thing that would be my default setting in many scenarios

But what I really want to know is how they will deal with wildlife - I'm extra careful and will brake / slow for birds hares deer etc, but what will the autonomous car or lorry do, will their be choosable setting ranging from , don't give a bugger to super careful ?

sorry - can't follow link - I only have 0.3mbs at the moment - so text only.[/QUOTE]
Good point. I'd never considered that (actually I hadn't concerned my grey matter with self driving cars at all...but you get my drift)

 

[Yellow Saddle]

I doubt that the image analysis will be so precise as to differentiate between an adult pushing a child in a pushchair and an adult pushing an adult in a wheelchair.

Of course not. "Subject X pushing a perambultor containing subject juvenile subject Y traversing on an adjacent trajectory to autonomous transport pod travelling on converse vector was deemed by subroutine A/X23 in algorithm XYZ to be present in lower hierarchy than fat man on bridge and hence was selected for instantaneous elimination."

Jokes aside, there are professional philosophers thinking about the problems as we speak. Even better, there are people employed to think just about the methods experts should use to think about the issues.

 

[classic33]

Of course, not, it will be in political speak.

Will the car understand that though?

 

[Yellow Saddle]

Will the car understand that though?

Yes. It will be translated for the car into machine language and hard-coded into its brain.

 

[classic33]

Yes. It will be translated for the car into machine language and hard-coded into its brain.

Google seems to favouring "cloud storage", so all we need is a power outage to really test it.

Look at where satnavs have "taken" some people already.

 

[Vapin' Joe]

Give the car another scenario. It rounds a corner, there's a tanker parked up on one side of the road, a fairly solid wall on the other. There's two cyclists who have moved out, that's why the tanker is stationary, to pass the tanker.

What does the car do? Save its occupants by hitting the wall, or hit the cyclists.

Simple. The car will go round the corner at a speed where it can stop in the distance it sees to be clear. Therefore it will neither hit the tanker nor have to swerve to avoid it.