Announcements: Universe of the Month! » Finding Universes to Join (and making yours more visible!) » Guide To Universes On RPG » Starter Locations & Prompts for Newbies » RPG Chat — the official app » USERNAME CHANGES » Suggestions & Requests: THE MASTER THREAD »

Latest Discussions: Life in the 21st century. » Song of the Runes » Plato’s Beard » Clues » Nihilism » Strange Tales From Hadean » Art Gulag [ Come get this Commish! ] » Visibility of Private Universes & Profile Customisation » Presuppositionalism » Aphantasia » Skill Trees - Good, Bad & Ugly » In-Game Gods & Gameplay Impact » Cunningham's Law » The Tribalism of Religion » Lost Library » Game Theory » The Hidden Void » Removing CS From an Indy Universe : Solution » On the Matter of New Players and Orphaned Plays » STOP BLAMING US FOR RPG BEING SLOW! »

Players Wanted: 18+ Writing Partner [Fantasy, Romance, Etc.] » 18+, Multi-Para to Novella Writers please! » Looking for roleplayers » Fun tale full of angels, demons, and humans » Looking for roleplayers » A Fairytale World in Need of Heroes & Villains! » Are You a Crime Addict? » Wuxia RP » Looking for roleplayers » New Realistic Roleplay - Small World Life ٩( ´・ш・)و » Mentors Wanted » MV Recruiting Drive: sci-fi players wanted! » Veilbrand: The Revolution » Gonna do this anyway. » Looking for Kamen Rider Players » Elysium » Looking for roleplayers for a dystopian past! » Revamping Fantasy Adventure RPG, need new players » Here, At the End of the World (supernatural mystery drama) » Seeking Role Players for a TOG Based RP »

Thought Experiment: Self-Driving Cars

a topic in Discussion & Debate, a part of the RPG forum.

Moderators: dealing with it, Ambassadors

Talk about philosophy, politics, news & current events, or any other subject you're interested in!

Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby VindicatedPurpose on Fri Apr 22, 2016 9:27 pm

So I was watching a video on TED. It was a thought experiment about self-driving cars.

So the gist of it was something like this...

Patrick Lin wrote:You are driving in a self-driving car in the near future. You are speeding along on a highway and you find yourself boxed in on all sides. Suddenly, a very large and heavy object falls off from the truck in front of you. Your car can not stop in time to avoid the collision, but it still needs to make a decision. It can continue straight and hit the object, swerve left and hit an SUV, or swerve right and hit a motorcycle. What should it prioritize? Your safety by hitting the motorcycle?Minimizing danger to others by not swerving, thus hitting the large object? Or take the middle ground by hitting the SUV (with its high safety rating)?


Here's a link to the video, in case you wanted to check it out.

https://www.youtube.com/watch?v=ixIoDYVfKA0

So, what's your take?
Like a stranger on a grate, or a skylark, or a taper, flying ever upward and knowing of love's satiety. Our dreams beyond the Sun and into the expanse of Night doth sound a quiet hymn.

Tip jar: the author of this post has received 0.00 INK in return for their work.

User avatar
VindicatedPurpose
Member for 10 years
Contributor Promethean Author Conversation Starter Inspiration World Builder Conversationalist Novelist Completionist Arc Warden Party Starter Beta Tester Greeter Visual Appeal Lifegiver Tipworthy

Re: Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby Aniihya on Sat Apr 23, 2016 11:16 pm

Why the hell would self-driving cars tailgate? I'd rather drive myself anyways.
Everybody! Unless you have been in a roleplay with me in the past and were loyal to it, do not PM or text me about joining your RP.

I do NOT do Pokemon, Yugioh, WoW or any such RPs.

Please be aware of this.

Tip jar: the author of this post has received 0.00 INK in return for their work.

User avatar
Aniihya
Member for 12 years
Promethean Conversation Starter Author Inspiration World Builder Conversationalist Friendly Beginnings Novelist Greeter Tipworthy Tipworthy Visual Appeal Person of Interest Lifegiver

Re: Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby VindicatedPurpose on Sun Apr 24, 2016 12:18 am

Well, nobody said you were tailgating, just that you were driving so fast that you wouldn't be able to slow down in time to avoid the large object. Also, you'll have to entertain the thought of being in a self-driving car for this thought experiment to work. So this is one of those few times where you should think inside the box.

Tip jar: the author of this post has received 0.00 INK in return for their work.

User avatar
VindicatedPurpose
Member for 10 years
Contributor Promethean Author Conversation Starter Inspiration World Builder Conversationalist Novelist Completionist Arc Warden Party Starter Beta Tester Greeter Visual Appeal Lifegiver Tipworthy

Re: Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby Aniihya on Sun Apr 24, 2016 10:24 am

"Boxed in means there is not even a cars length to maneuver around others. That is tailgating, due to the lack of distance to anyone. And a self-driving car would react according to which outcome logically causes the least deaths and damages, if programmed to do so. Otherwise it could be a cheap ass piece of American engineering and take the least logical conclusion.

Tip jar: the author of this post has received 0.00 INK in return for their work.

User avatar
Aniihya
Member for 12 years
Promethean Conversation Starter Author Inspiration World Builder Conversationalist Friendly Beginnings Novelist Greeter Tipworthy Tipworthy Visual Appeal Person of Interest Lifegiver

Re: Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby VindicatedPurpose on Sun Apr 24, 2016 11:59 am

On the contrary boxed in is boxed in and tailgating is tailgating, but that's not the point we're here to discuss.

You say that it would react according to "which outcome logically causes the least deaths and damages, if programmed to do so."

So do you mean that we should program it according to that particular standard which causes the least deaths and damages?

Or are you questioning whether it is programmed to do so? Because it looks like there isn't any other standard by which we should program a car according to that statement.

If the motorcycle had to 2 people on it, and you were the only person in your car, it would drive forward and kill you because by that standard the least deaths will have been achieved.

So are we to say that we'll just neglect your own safety if it means more lives will be saved?

Wasn't the purpose of the self-driving car to make driving safer? But suddenly now if you're not carpooling, your car will default to killing you on the open road. My my, what dystopian times await us.

Tip jar: the author of this post has received 0.00 INK in return for their work.

User avatar
VindicatedPurpose
Member for 10 years
Contributor Promethean Author Conversation Starter Inspiration World Builder Conversationalist Novelist Completionist Arc Warden Party Starter Beta Tester Greeter Visual Appeal Lifegiver Tipworthy

Re: Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby FyreT1ger on Sun Apr 24, 2016 4:41 pm

If you're going on the basis of logic and safety, there is a higher chance of survival by you and other parties by hitting either another car or a large object while inside a car. Hitting a motorcycle, bicycle, or scooter will most likely kill the cyclist or scooter rider/driver no matter how another motorist or smart car tries to avoid it.

On the upside, the cycles and scooter are more maneuverable than a regular/smart car. If you add human behavior to the equation, the (motor)cyclist will move his or her vehicle out of range first and your smart car can follow. On a highway with the large object falling from a moving vehicle, it may not fall straight back so the cyclist will want to get the heck out of the way, because as I said before, motorcyles and bicycles getting hit have a very low chance of survival for any riders aboard.

That's my either over-complicated or exceedingly practical response.

Tip jar: the author of this post has received 0.00 INK in return for their work.

User avatar
FyreT1ger
Member for 12 years
Conversation Starter Author Conversationalist Friendly Beginnings Novelist Donated! Contributor Lifegiver

Re: Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby Aniihya on Mon Apr 25, 2016 8:57 am

VP, do you see any questions? I made statements indicated by a "would be" scenario. And safer driving doesn't mean to just protect the passenger of said car at all costs. It would also mean that your vehicle doesn't endanger other motorists.

And now you are jumping to conclusions when I am trying to express my view as clearly as possible.

Plus would the large object be a spindle, another car, a case full of knives, a forklift? Like FyreT1ger indicated, it depends on the object too as not all objects react the same to hitting the ground. If it falls differently, it could hit the SUV or the motorcycle or just bounce over you.

Tip jar: the author of this post has received 0.00 INK in return for their work.

User avatar
Aniihya
Member for 12 years
Promethean Conversation Starter Author Inspiration World Builder Conversationalist Friendly Beginnings Novelist Greeter Tipworthy Tipworthy Visual Appeal Person of Interest Lifegiver

Re: Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby VindicatedPurpose on Mon Apr 25, 2016 4:34 pm

I see plenty of questions, actually. Don't get me wrong, I'm not saying your statements are questions. Those are general questions that I thought of and posed not specifically to anyone.

You say safer driving doesn't just mean protecting the passenger of the car at all costs, and that it would not endanger other motorists. Theoretically, and in an ideal world, that's what we all want, and I don't disagree with that. The question to this is is that possible?

The experiment wants to go with the idea that it's not, that somebody will or may die. We have seen throughout the history of motorized transportation that there have been moments like the above situation, where they can be mutually exclusive.

We're talking about a computer built by humans, if we are fallible, how do we guarantee that computers are infallible? Or is it even possible to make such a guarantee?

If you add human behavior to the equation, the (motor)cyclist will move his or her vehicle out of range first and your smart car can follow.


That's one possibility in a range of possibilities. When you're moving that fast, things happen in seconds, the biker's reaction time might not be on par with what we'd want.

Based off of what the video outlined, I was going off the assumption that the large object was capable of killing you if you were to continue straight on. That's probably the route they would have wanted you to take. You guys did check out the link right?

Then again, you could frame the question as this, who would you kill? And was it justified? In multiple other iterations, it might not even be a van on your left flank, it could be another motorist without a helmet.

Thoughts?

Tip jar: the author of this post has received 0.00 INK in return for their work.

User avatar
VindicatedPurpose
Member for 10 years
Contributor Promethean Author Conversation Starter Inspiration World Builder Conversationalist Novelist Completionist Arc Warden Party Starter Beta Tester Greeter Visual Appeal Lifegiver Tipworthy

Re: Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby Aniihya on Tue Apr 26, 2016 1:41 am

Why can't you ask something simpler like the test with the train tracks and the lever or the train tracks and whether or no you should throw a fat man in front of a train to save a group of people stuck on the tracks?

Tip jar: the author of this post has received 0.00 INK in return for their work.

User avatar
Aniihya
Member for 12 years
Promethean Conversation Starter Author Inspiration World Builder Conversationalist Friendly Beginnings Novelist Greeter Tipworthy Tipworthy Visual Appeal Person of Interest Lifegiver

Re: Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby VindicatedPurpose on Thu Apr 28, 2016 10:38 pm

Lol Aniihya you know why I can't ask that.

1) It's been asked before. Admit it.

2) This question has a relevant and specific context, and that's not to say that the other question doesn't. It might even bring up different possibilities because of the context.

I mean, let's face it, in the next couple of years we'll see more electric vehicles, smart cars, and virtual reality. We're getting more and more plugged in, and at the apex of this is artificial intelligence. Where we must ask ourselves even more questions.

Someone once told me that new technology solves old problems by presenting new ones.

For example, a teleporter may cut your commute from Los Angeles to New York down to just 2 minutes, but you might end up creating temporal copies of yourself that will have to be destroyed lest you disturb the space-time continuum. My science is fast and loose here, but like they say if we can imagine it, we can...yada yada, you know the rest.

Tip jar: the author of this post has received 0.00 INK in return for their work.

User avatar
VindicatedPurpose
Member for 10 years
Contributor Promethean Author Conversation Starter Inspiration World Builder Conversationalist Novelist Completionist Arc Warden Party Starter Beta Tester Greeter Visual Appeal Lifegiver Tipworthy

Re: Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby Aniihya on Fri Apr 29, 2016 5:37 pm

1) There are five variants of it.

2) Doesn't seem specific enough to me.

I am always thinking and always solving math problems in my head. I do not go too far into science fiction when talking about the real world. Plus stuff we imagine ends up being totally different in the future (see 1960s "WORLD OF TOMORROW!" stuff). Scientific progress is based on will and goals. You make steps and not jumps.

Tip jar: the author of this post has received 0.00 INK in return for their work.

User avatar
Aniihya
Member for 12 years
Promethean Conversation Starter Author Inspiration World Builder Conversationalist Friendly Beginnings Novelist Greeter Tipworthy Tipworthy Visual Appeal Person of Interest Lifegiver

Re: Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby VindicatedPurpose on Sat Apr 30, 2016 3:05 pm

Sure, and there's a reason for five variants of it.

If this context isn't specific enough, you can debate about it in another thread.

The 1960s' World of Tomorrow highlighted one among a million possibilities of what the world would look like. Fact is fiction until it is proven. Fiction is fiction until it becomes reality. We dream of it, it can happen.

When I mentioned artificial intelligence, it's not even science fiction anymore. It was fiction maybe some thirty or forty years ago when Stanley Kubrick was talking to Steven Spielberg about a film idea. Now? Not so much. We've got robotic art curators and Facebook's whole algorithm of finding things that match your interests is building an avatar and profiling you.


The case here is a car where someone programs into it a decision making algorithm, and despite the desire to maximize efficiency and safety, somehow somewhere, it will fail.

The real world is a combination of what has happened, what is happening, and what could happen.

Tip jar: the author of this post has received 0.00 INK in return for their work.

User avatar
VindicatedPurpose
Member for 10 years
Contributor Promethean Author Conversation Starter Inspiration World Builder Conversationalist Novelist Completionist Arc Warden Party Starter Beta Tester Greeter Visual Appeal Lifegiver Tipworthy

Re: Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby FyreT1ger on Mon May 09, 2016 8:08 am

"The case here is a car where someone programs into it a decision making algorithm, and despite the desire to maximize efficiency and safety, somehow somewhere, it will fail."

This is why I prefer dealing with humans over machines. Humans make mistakes yes. Some humans are mean-spirited while others are nice. Some talk too much and others are quiet. Some humans are annoying, and some are not. Human beings have different personalities, but one thing they all have in common is that they still have the ability to learn and change. Machines can't change, and can't do anything whatsoever without a power source, such as electricity. When a storm cuts out the electric lines and the power plant all of those machines are dead in the water, as well as all the effort that went into the algorithms and whatever else is needed for the machine to work.

Earlier you said, "So this is one of those few times where you should think inside the box." I can't stay inside a box in these no-win situations. Same goes for the train track question Aniihya references. I actually was walking on a train track as a train was coming and I see perfectly well that I can walk away from that train track without getting killed. I saw the train coming and simply picked up the pace and was away from the train track before and bars came down blocking the track.

Where I walked that day a track crossed over the road, so it didn't have the high walls on the side as it would in a station stop, but those high walls are there so NO ONE ends up on the track by accident, which would logically limit the danger anyway. Train tracks also generally have fences beside them as well, except as in the case of the road. A fence in the middle of the road would impede general vehicle traffic.

Well I apparently went on a tangent, but I did technically already answer this question and don't really have another response anyway. I'd still want to follow a smaller vehicle out of a dangerous situation.

Tip jar: the author of this post has received 0.00 INK in return for their work.

User avatar
FyreT1ger
Member for 12 years
Conversation Starter Author Conversationalist Friendly Beginnings Novelist Donated! Contributor Lifegiver

Re: Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby Sara Whitley on Wed May 18, 2016 3:39 pm

From the perspective of the rider, go for the SUV. As the rider you are aware that there is in fact an SUV there and that would be the best scenario, hopefully protecting yourself and not endangering others too badly.

But of course, the car is self-driving. It was programmed, and realistically, it would not be programmed to go for the SUV or the motorcycle. Doing so would be counter intuitive to the collision avoidance systems and programming in the capacity to do that might cause huge legal and insurance implications.

So if the computer was programmed from the company, it probably would continue to drive, if speeding up or slowing down would not stop the collision.

Aside from this it's kind of hard to do a thought experiment from the perspective of the rider, who's thoughts would be irrelevant.


A computer was once programmed to play Tetris. Well it didn't do it so well, and when it reached a point where it could no longer play, it paused the game--forever. It was the only remaining option for it to not lose.
Thinking inside the box, it sounds like both the AI and the human intelligence are both kind of screwed.
Creator of the Fantasy Sandbox roleplay.

Tip jar: the author of this post has received 0.00 INK in return for their work.

Sara Whitley
Member for 11 years
Conversation Starter Author Conversationalist Arc Warden Lifegiver Tipworthy Visual Appeal

Re: Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby VindicatedPurpose on Sat May 21, 2016 12:10 pm

@FyreT1ger

Unfortunately, humans will be the ones coding the programs that will run the cars. I think the best you can do with this experiment is to think inside and outside the box at the same time.

Pick a path, follow it, and think about it; and then criticize the thought experiment. That's how I've managed to stay sane. Ha.

@Sara

Screwed, indeed. You're right, it's not about the rider, but about which direction the computer algorithms choose to go. If it's a lose lose situation, I don't think this computer will have the same luxury as the Tetris program. By this point, we're just playing the game not to lose, or to not lose too much.

The best part out of all of this is that it's just a thought experiment. We have the luxury to deliberate before it becomes a possible reality. So I'm just glad for that.

Tip jar: the author of this post has received 0.00 INK in return for their work.

User avatar
VindicatedPurpose
Member for 10 years
Contributor Promethean Author Conversation Starter Inspiration World Builder Conversationalist Novelist Completionist Arc Warden Party Starter Beta Tester Greeter Visual Appeal Lifegiver Tipworthy

Re: Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby Spearheader on Mon Jul 25, 2016 9:34 am

VindicatedPurpose wrote:On the contrary boxed in is boxed in and tailgating is tailgating, but that's not the point we're here to discuss.

You say that it would react according to "which outcome logically causes the least deaths and damages, if programmed to do so."

So do you mean that we should program it according to that particular standard which causes the least deaths and damages?

Or are you questioning whether it is programmed to do so? Because it looks like there isn't any other standard by which we should program a car according to that statement.

If the motorcycle had to 2 people on it, and you were the only person in your car, it would drive forward and kill you because by that standard the least deaths will have been achieved.

So are we to say that we'll just neglect your own safety if it means more lives will be saved?

Wasn't the purpose of the self-driving car to make driving safer? But suddenly now if you're not carpooling, your car will default to killing you on the open road. My my, what dystopian times await us.


In response to the bolded: No. It will only kill you if you're in a situation where some death occurring is likely. This is obviously not going to happen all the time.

I would actually have it ram into the object, so long as the hospitals would just let me die after a certain point.

Tip jar: the author of this post has received 0.00 INK in return for their work.

User avatar
Spearheader
Member for 5 years
Conversation Starter Author Lifegiver

Re: Thought Experiment: Self-Driving Cars

Tips: 0.00 INK Postby philosophotter on Wed Sep 14, 2016 12:07 pm

It seems like hitting the SUV would be best choice only because that gives the chance for the least loss of life and utilitarianism is totally a thing people instinctively gravitate to the the collective. If the car hit the object, that's pretty much a guaranteed death, but if the SUV was hit, then the blow would be glancing (hopefully) and would result in more spinning and a more gradual stop that just....car. Object. Boom. Death.
Image

Tip jar: the author of this post has received 0.00 INK in return for their work.

User avatar
philosophotter
Member for 5 years
Conversationalist


Post a reply

Make a Donation

$

Become a Patron!

RPG relies exclusively on user donations to support the platform.

Donors earn the "Contributor" achievement and are permanently recognized in the credits. Consider donating today!

 

Who is online

Users browsing this forum: No registered users and 1 guest