Philosophers are building ethical algorithms to help control self-driving cars

[Kukielka]

3 hours ago, mr moose said:

I won't get in a car (or any man made tool) that doesn't prioritize my life over everyone else's.

 

That's not what I was talking about.

Say you are save, however your breaks fail and you have to choose, run over 5 kids, or run over 5 elderly?

You cant possible give an "ultimate" answer thats always right.

[Canada EH]

5 elderly is the right answer

[Kukielka]

Just now, Canada EH said:

5 elderly is the right answer

I hope that you are aware that there is no "right" answer.

Cause if you actually think that, that's very slim minded.

[mr moose]

2 hours ago, AshleyAshes said:

I got some troubling news about air travel for you then...

As far as  I know all air travel technology prioritizes the safety of it's passengers.  Suffice to say air travel scares the shit out of me and I won't do it unless I have to.

1 minute ago, Kukielka said:

That's not what I was talking about.

Say you are save, however your breaks fail and you have to choose, run over 5 kids, or run over 5 elderly?

You cant possible give an "ultimate" answer thats always right.

Then don't make it decide beyond the safety of the occupants.  As I said in the 5th post, it doesn't matter what decision the AI makes there will always be people who think it's wrong.  

[Kukielka]

Just now, mr moose said:

As far as  I know all air travel technology prioritizes the safety of it's passengers.  Suffice to say air travel scares the shit out of me and I won't do it unless I have to.

Then don't make it decide beyond the safety of the occupants.  As I said in the 5th post, it doesn't matter what decision the AI makes there will always be people who think it's wrong.  

So if the AI seems you are safe, then all other safety mechanisms should kick in?

Say the breaks fail, say there are people on the street you could dodge,

Following your "logic": Run over these people, atleast I am safe.

 

Kinda dumb, don't you think?

[mr moose]

Just now, Kukielka said:

So if the AI seems you are safe, then all other safety mechanisms should kick in?

Say the breaks fail, say there are people on the street you could dodge,

Following your "logic": Run over these people, atleast I am safe.

 

Kinda dumb, don't you think?

You're adding unnecessary conditions.  If the occupants are safe and the car can avoid people then it should. If the car can't avoid people then it should take the course of least danger to the occupants.

 

 

 

[BlueChinchillaEatingDorito]

Would driverless cars sound horns at each other then.

[AshleyAshes]

11 minutes ago, mr moose said:

As far as  I know all air travel technology prioritizes the safety of it's passengers.  Suffice to say air travel scares the shit out of me and I won't do it unless I have to.

Funny story about Post 9/11 air travel; If your aircraft is ever hijacked and the passengers are unable to regain control of the air craft the Air Force will kill you.

 

Even on the day of 9/11 itself, there were commercial airliners that had problems responding to air traffic control, in one case unarmed Air National Guard F-16's were deployed on a potential kamikaze mission to intercept of of the airliners.  Thankfully the communications issues involving these aircraft was resolved and the aircraft had not been taken.  There have been multiple cases since of commercial aircraft which have failed to response to ground communications which have been intercepted by air forces and escorted until they responded again.

 

Quit simply put, the entire 'system' involving air travel, in the current climate, is that if your air craft stops responding and does not resume communication once intercepted, and you are approaching a populated area, they will kill you.  The lives of the passengers are a lower priority than the lives of those on the ground.  Good motivation for the passengers to retake the cabin at all costs; Because that's the only way you will touch the earth again alive.  Also, this is why 9/11 style hijackings are not repeated.  The genie is out of the bottle and the passengers WILL revolt at all costs because it's the only scenario that doesn't guarantee their death and as a result such an attempted attack has an exceptionally high chance of failure.

[Kukielka]

Just now, mr moose said:

You're adding unnecessary conditions.  If the occupants are safe and the car can avoid people then it should. If the car can't avoid people then it should take the course of least danger to the occupants.

See? Now you'r exactly doing what the "quiz" is asking for, you create priorities and want the ai to behave like that according to your moral compass.

Saying this is pointless, is just invalid.

[mr moose]

1 minute ago, AshleyAshes said:

Funny story about Post 9/11 air travel; If your aircraft is ever hijacked and the passengers are unable to regain control of the air craft the Air Force will kill you.

 

Even on the day of 9/11 itself, there were commercial airliners that had problems responding to air traffic control, in one case unarmed Air National Guard F-16's were deployed on a potential kamikaze mission to intercept of of the airliners.  Thankfully the communications issues involving these aircraft was resolved and the aircraft had not been taken.  There have been multiple cases since of commercial aircraft which have failed to response to ground communications which have been intercepted by air forces and escorted until they responded again.

 

Quit simply put, the entire 'system' involving air travel, in the current climate, is that if your air craft stops responding and does not resume communication once intercepted, and you are approaching a populated area, they will kill you.  The lives of the passengers are a lower priority than the lives of those on the ground.  Good motivation for the passengers to retake the cabin at all costs; Because that's the only way you will touch the earth again alive.  Also, this is why 9/11 style hijackings are not repeated.  The genie is out of the bottle and the passengers WILL revolt at all costs because it's the only scenario that doesn't guarantee their death.

The airforce shooting you down is not the same as AI or some intrinsic part of the tech on a plane dumping you in the side of hill instead of attempting to land for some other fear.  The airforce shooting you down is a complete separate entity that no one but the airforce has control over.

Just now, Kukielka said:

See? Now you'r exactly doing what the "quiz" is asking for, you create priorities and want the ai to behave like that according to your moral compass.

Saying this is pointless, is just invalid.

What?   If it manages to discern a moral method for resolving the trolley problem that is unanimously accepted,  then I will concur it wasn't pointless.  But given people have been trying to solve said problem for 1000's of years, then I don't see how asking people the same questions again is going to make any difference other than to garner the same responses and come to the same conclusions.  I.E There is no unanimously agreed correct answer,  there has never been one so why would asking again change anything?

[AresKrieger]

On 2/13/2018 at 12:45 AM, GabD said:

a philosophy degree is apparently still going to be relevant for a while.

It was never relevant to begin with, this only demonstrates that further, no one will buy a car designed to disregard their safety if some idiot runs in the road, additionally the statement below raises extreme red flags.

Quote

It’s not just about how many people die but which people die or whose lives are saved,” says Evans. It’s possible that two scenarios will save equal numbers of lives, but not of the same people.

 

 

I do not want a moral car, I want one who values me over all else and though most others will not say it I doubt that a car that could so easily disregard its passengers will sell well, instead of wasting time on this nonsense the makers of these cars should simply focus on improving the braking reaction of said cars rather than driving them off a cliff at the first sign of some dumb kid in the street.

[mr moose]

2 minutes ago, AresKrieger said:

It was never relevant to begin with, this only demonstrates that further, no one will buy a car designed to disregard their safety if some idiot runs in the road, additionally the statement below raises extreme red flags.

 

 

I do not want a moral car, I want one who values me over all else and though most others will not say it I doubt that a car that could so easily disregard its passengers will sell well, instead of wasting time on this nonsense the makers of these cars should simply focus on improving the braking reaction of said cars rather than driving them off a cliff at the first sign of some dumb kid in the street.

Fortunately most philosophers know there is no real answer and they are most likely more interested in making sure some dimwit with a marketing degree doesn't turn AI into a soylent green issue.  

 

I say this because most of the philosophers I have either read or talked to are more interested in providing a moral framework for discussion that goes into the making of laws and governmental policies.   I have yet to meet or read about one who fervently has argued a point because they believe it to be true above all other points.   Not that they aren't out there, just that their poor philosophical skills lowers them on the relevance ladder so we only here about them in supermarket tabloids and conspiracy websites.

[Techstorm970]

On 2/13/2018 at 3:05 AM, VegetableStu said:

"Good morning Mr. Bond. How would you like to Die Another Day?"

FTFY

[Canada EH]

58 minutes ago, Kukielka said:

.

well 5 children who have 80 yrs left in their life, or 5 elderly who have 5 yrs left in their life

its a no brainer fella

 

your rational would be, dont cover your child at the vegas concert as madman shooting from hotel room

no no

cover your elderly parents

 

makes no sense

[GabD]

On 13/02/2018 at 5:46 AM, Whiskers said:

@GabD Please update your thread to follow the Tech News & Reviews Posting Guidelines.

 

Edit: Oh, and welcome to the community.

I did read the guidelines before posting.  I don't know what to do then.
The post a) starts with my own input, b) links to a relatively reputable source, c) quotes important bits but not the whole article, and d) makes use of quotation marks for all non-original text.

Is it too much of a wall of text? Should I use a full-on quote box instead of just the quotation marks?

 

On 13/02/2018 at 3:55 PM, SpaceGhostC2C said:

However, we cannot make informed decisions if we don't understand the consequences of our choices. This team puts together philosophers with deep knowledge of the ethical systems that guide much of our legislation with engineers that can code those principles into self-driving algorithms. Simulating the behavior of a car that follows that algorithm in a wide range of situations is necessary to avoid passing legislation based on what some dude (or army of dudes for that matter) on an armchair cleverly conceived was a good idea. The point is not for some philosopher to tell us "I determined that the best ethics is X, now go do it", but precisely the opposite: to gain understanding of what real-life situations would look like under different systems.

This.

 

On 13/02/2018 at 5:43 PM, Sauron said:

You still gave no compelling reason for which they should be in any way more qualitied to do this than anyone else, or why they need half a million to do it, unless it's just their professor paycheck for that time frame.

@SauronI'll attempt an answer to this as such : it is a matter of levels of discourse. (Tl;dr version at the end)

Academics in the field of Philosophy are specialised in the argumentation and justification of theses. A philosophical discourse is a discourse on a discourse; such as "how we think about things", and what differentiates sound reasoning from unsound reasoning.
"Philosophers" (and by that I generally mean academic experts, not most self-titled writers) are effectively more qualified to deal with matters pertaining to this specific field.
Even though pretty much all humans have to do a bit of this in order to function, and are as such at least a bit familiar with the principles of reasoning, professors of Philosophy are the people whose job it is to be experts about it. Like how a biologist's job is to be an expert in biology (or rather, in a specific branch of biology). Similarly, professors of Philosophy are specialised in specific branches of the discipline.

Explaining it further requires explaining the difference between contemporary Philosophy and the contemporary Sciences (but I won't touch on why we stopped calling the Sciences "natural philosophy" and stopped calling Philosophy "science" singular).

Philosophy exists in academia as a meta-scientific field. There are a few other meta-scientific fields, which all cover specialized types of inquiry.

Academics in the field of Philosophy are experts in branches of that field, which include (not exhaustively) the branches of Logic, Epistemology (the study of knowledge and how to obtain it), Ontology (how to differentiate things that are and things that aren't) and Ethics. As such, Philosophy is a specialised meta-scientific field which is tasked with the stuff that directs and oversees the development of Science (among other things).

 

While many professors of philosophy are philologists (interpretative historians whose work pertains to the History of ideas), many also do active research pertaining to contemporary subjects, such as "how can we know that our method for figuring out «whether the Large Hadron Collider poses risks or not» will be reliable".
As such, it is thanks to (and because of) epistemologists that we can be fairly certain that "we are correct in believing that the Large Hadron Collider is safe to use within the parameters that have been set" is true.

From my (imperfect) point of view (as an undergrad student), contemporary ethics works a lot like epistemology. It is also about "how can we know what we want to know", and "how can we know when/if we know what we want to know"; in which case the "knowing" is about "what choice(s) would best be made in X context". (Deontologists basically just take X as a constant more often than not, so I'll speak from the context-sensitive angle since it covers everything.)

If your field has terms like "best practice" or "deontological code", that's the product of prescriptive reasoning, not unlike the products of research in Ethics (and it sometimes directly IS the product of the work of "philosophers").

Tl;dr: Ethicists are academic experts in providing (usually context-sensitive) frameworks for prescriptive reasoning. This makes that work both relevant and important for many, many things related to a) scientific research, b) technological development, c) legitimate uses of power, d) solving problems, etc.

As for the half-million dollard in funding:
Evans has won a three-year, $556,650 National Science Foundation grant to construct ethical answers to questions about autonomous vehicles (AVs), translate them into decision-making algorithms for AVs and then test the public health effects of those algorithms under different risk scenarios using computer modeling. 

He will be working with two fellow UML faculty members: Heidi Furey, a lecturer in the Philosophy Department, and Asst. Prof. of Civil Engineering Yuanchang Xie, who specializes in transportation engineering. The research team also includes Ryan Jenkins, an assistant professor of philosophy at California Polytechnic State University, and experts in public health modeling at Gryphon Scientific."

Source: https://www.uml.edu/News/stories/2017/SelfDrivingCars.aspx

[GabD]

On 14/02/2018 at 7:13 PM, mr moose said:

There is no unanimously agreed correct answer,  there has never been one so why would asking again change anything?

Having people agree on things isn't the immediate point of this. People's differing opinions are also partly caused by many things other than rational justification. Here the point is finding the most rationally justified answer(s). The part about convincing people to change their mind is a whole different thing.

- We have no reason for believing that finding a "most justified answer" would somehow be impossible. The principles of rationality aren't exactly subjective.
- We only have reason to doubt that people would initially unanimously accept it, if there were to be such an answer. No reason to dismiss the very possibility of it being widely accepted.

Also, this is not at all "asking again". It's not even "asking people".

[GabD]

On 14/02/2018 at 7:32 PM, AresKrieger said:

It was never relevant to begin with, this only demonstrates that further, no one will buy a car designed to disregard their safety if some idiot runs in the road, additionally the statement below raises extreme red flags.

Now, we need to differentiate "action which can be reasonably justified as more likely/guaranteed to improve the state of things in some way" and "action which is more popular". They aren't always the same, even though they sometimes are.

In most countries, AV heuristics will be at least set guidelines via legislation. Research such as this is funded by entities like the National Science Foundation to figure out what would be the best practices in AI heuristics. I.e. the goal is to help solve potential or existing problems.

Automobile manufacturers are those who specifically care about what sort of products individual consumers want. But "what consumers want to buy" may have to be restricted by "what can justifiably be available".

I mean, there are reasons why individual consumers aren't allowed to buy artillery ordinance for home defense, and still would not be allowed to, even if that's what all consumers wanted.
This is an exaggeration, mind you.
But the point is that somewhere along the line a prescriptive reasoning is made, and that's "doing ethics", regardless of wether or not the title of the person(s) doing that reasoning is "philosopher", "ethician", "judge", "senator", "citizen" or whatever else.

I mentioned the relevancy of a degree in Philosophy in this context because it is specifically meant to train people in the methods of justification and rational discourse, with regard to discourse on facts about the world (scientific), and discourse on the way we discourse about facts about the world (meta-scientific).

[Sauron]

 @GabD while all that is informative, I still find it a little hard to believe their potential contribution in this particular area doesn't overlap with the engineers' own experience in creating reliable systems and accounting for different scenarios.

[GoldenLag]

On 15.2.2018 at 1:45 AM, Canada EH said:

well 5 children who have 80 yrs left in their life, or 5 elderly who have 5 yrs left in their life

its a no brainer fella

 

your rational would be, dont cover your child at the vegas concert as madman shooting from hotel room

no no

cover your elderly parents

 

makes no sense

How many elderly people outweigh one child then? Its not allways a no-brainer. How about a driver and two children vs a crowd of elderly people vs 5 middleaged people. Udk what senarios this situation would occur. Which one will the AI deem worthy of survival and which group of people will have to die. Its not often you get a no-brainer situation

[Whiskers]

6 hours ago, GabD said:

I did read the guidelines before posting.  I don't know what to do then.
The post a) starts with my own input, b) links to a relatively reputable source, c) quotes important bits but not the whole article, and d) makes use of quotation marks for all non-original text.

Is it too much of a wall of text? Should I use a full-on quote box instead of just the quotation marks?

 

Ah, fair enough. The personal opinion part came across more as a summary of the article's contents rather than an opinion. In that case, don't worry about it.

[The Benjamins]

On 2/13/2018 at 3:35 AM, Sauron said:

The trolley problem isn't actually a real thing. There are always more variables and more options.

This is pretty ridiculous, it hardly takes half a million $ to run simulations on existing frameworks and quite frankly the philosophers' input seems pretty superflous. In the end, a perfect trolley problem (which doesn't exist in the real world, but let's suppose it did) does not have a fixed solution by its very nature.

 

I also find it extremely ironic that these philosophers believe they have the authority to make this decision for others. One of the first thing you learn in philosophy courses is that being a philosopher does not make you more right than others. Realistically, the way it's most likely going to pan out is that governments will introduce a law that values the life of most over the life of a few every time and the cars will have to abide by it, and this "research" will have been just a waste of money and time.

agree with this, the way I see it is the car should protect passenger first, then minimize external damage. (hitting less objects)

 

Time spent on picking what to hit is time wasted to prevent hitting it in the first place. The correct answer to the autonomous car trolley question is to have it hit no one.

[DildorTheDecent]

On 13/02/2018 at 8:40 AM, mr moose said:

  Harambe anyone?

Too soon.

 

Mods pls ban.

[Syntaxvgm]

6 minutes ago, DildorTheDecent said:

Too soon.

 

Mods pls ban.

The only thing deader than that meme is Harambe.

[The Benjamins]

7 minutes ago, DildorTheDecent said:

Too soon.

 

Mods pls ban.

What do you mean too soon? that topic is 100% relevant to this due to it being a perfect real world example of the trolley problem.

[DildorTheDecent]

3 minutes ago, The Benjamins said:

What do you mean too soon? that topic is 100% relevant to this due to it being a perfect real world example of the trolley problem.

*whoosh* well there it goes...the meme, right over your head...

 

Someday you'll understand.