The Ethical AI Question Of Whether Self-Driving Cars Ought To Be A Good Samaritan And Forewarn When Human-Driven Cars Are Going To Crash Into Each Other

Mind your own business.

That’s a common refrain about not sticking your nose into the activities of others.

Extend a helping hand.

That’s also a common refrain and refers to the notion of being helpful to others, even though they might be complete strangers and you do not know them at all.

Which of those two catchphrases or words of wisdom would you choose?

You probably make daily decisions about those two possibilities. There are situations and settings wherein you opt to mind your own business. At times, it might be quite tempting to step into the middle of something, but you weigh the pros and cons of doing so, and then at times move along and do not get into the fray. On the other hand, there are times that you decide it is best to jump into the swimming pool, as it were, and get engaged.

Let’s turn this somewhat conceptual or philosophical discussion into something very grounded and real.

I was driving my car the other day and had come up to an intersection to make a left turn. The left turn was controlled by a traffic light. The traffic light was red. I came to a proper and safe stop in that left turn lane or pocket and waited patiently for the traffic signal to show me a green arrow so that I could proceed to make the turn.

What do you do while waiting in such circumstances?

Some people seem to look down at their smartphones. This is not a good idea. You ought to keep your eyes and awareness on the roadway at all times. I realize that you could argue that since the vehicle is stopped in the proper place and waiting for the light to turn green, there is nothing particularly sinister or wrong about looking down at a cellphone. You aren’t going anywhere anyway, presumably.

The concern about your looking down at your smartphone is that you can become distracted. Your attention is no longer focused on the driving scene.

You’ve undoubtedly seen others that were caught looking down at their cellphones and then the light does go green. At this point, they are still not moving ahead, though they are supposed to do so. Cars waiting behind this driver will often get exasperated at the delay, sometimes honking their horn. This can readily and sadly escalate into a burst of road rage that is bad for everyone nearby.

Another outcome could be that when the light does turn green and you start ahead, you’ve now got an outdated comprehension of the driving environment. Suppose that a pedestrian has entered into the crosswalk and they are now nearing the front of your vehicle. The pedestrian is not supposed to be there in that they are improperly crossing when you have green. Nonetheless, there they are.

Because you were glancing down at your smartphone, you did not see the pedestrian as they gradually entered and walked along in the crosswalk. Your mental model of the driving scene was that the crosswalk was still clear and as soon as you got the green arrow, you could hit the gas and get going.

Unfortunately, you opted to hit the gas before reexamining the street scene. Your vehicle begins to move forward and all of sudden, now that your eyes are once again viewing the roadway, you see that a pedestrian has “magically” appeared in front of your car.

How in the heck did that person appear out of thin air?

Well, it wasn’t thin air, it was that you had a mental image or belief that was based on what you saw a few seconds earlier. You did not expect that a pedestrian would make a crossing. In theory, the crosswalk should have been entirely empty. Upon looking up, your focus became moving ahead, nearly out of habit and the assumption that the path forward should be ready for your use. That mental model got shaken up when you at the last moment detected the pedestrian.

I don’t want to belabor the point and merely point out that the driving scene is everchanging and the expectation is that drivers will remain attentive to the roadway at all times. Looking down at a smartphone or to adjust your radio, or perhaps reaching over to grab a burger and some fries that you purchased a few minutes earlier, all of those actions can divert your eyes and cause you to no longer have an up-to-date understanding of the driving scene.

Okay, let’s assume that you are going to keep your attention riveted to the roadway. Most of us do not do so, and it is altogether human that we don’t. Regrettably, a sizable portion of the annual 40,000 fatalities due to car crashes in the US, and the approximate 2.3 million injuries are due to distracted drivers. That’s why you see so many billboards and roadway signs that implore you to keep your eyes on the road, hoping that the number of deaths and injuries will be reduced.

What can you do if indeed your eyes are watching the roadway?

Some say it is boring to be gazing at the roadway.

Much of the time, such as while sitting and waiting to make a left turn, there is nothing notable that will be happening in the driving scene. Other cars will come and go. Pedestrians will come and go. Most of that will not have any bearing on your effort to make that left turn. Boredom arises.

You could say that this boredom is a good thing. It means that apparently nothing is occurring in the driving scene that pertains to your driving actions. You can rest easy, for a moment, and remain calm at the wheel.

Only if you suddenly see something untoward that might threaten your vehicle would you become suddenly sparked into heightened awareness. Imagine that another car comes along next to you and seems to be veering toward your vehicle. This is going to raise the hairs on the back of your neck. Thank goodness you were watching the roadway. By seeing this potential calamity, you can take evasive action.

Imagine though that you weren’t in any danger and yet you did see a dangerous situation brewing that could possibly harm others.

This happened to me as I was sitting in the left turn lane at the red arrow.

I could readily see that cars were coming in the opposing direction in the other lanes to my left on the other side of the divided street. At the far corner of this intersection, cars were coming up to the intersection and wanting to make a right turn. They were faced with a red light and had to come to a stop before proceeding.

A car that had earlier made that right turn was stopped just beyond the corner and was letting someone out of the vehicle. Perhaps it was a ridesharing service and the passenger wanted to get out at that location. The bad news was that this was a busy corner, and the corner was quite obscured by a bunch of trees that extended along the edge of the sidewalk.

In short, the car that was about to make the right turn was coming up fast to do so and would be unable to see around the corner, ergo not able to realize that a car was stopped just beyond there. If the driver making the right turn did so without fully stopping and gingerly creeping forward, they would likely go around the bend at a fast clip and find themselves shockingly about to hit another car (the one that was sitting still or semi-parked just around the corner).

Allow me to clarify that none of that was of any direct relevance to me, at least with respect to my car and my safety. I could see this happening as though it was occurring on a stage and these were actors in a play. If the car that was turning rammed into the other car that was semi-parked, it would not make a whit of difference to my situation. I was free and clear of any possible involvement in the calamity that was about to arise. This whole matter was merely within my driving scene in total, and I could see what was likely going to occur.

And this takes us back to my initial opening remarks.

Imagine that you were immersed in this driving setting and were the person driving my car. It was you that could see what was about to happen. Rather than looking down at your smartphone to check your text messages or watch a cat video, you were paying attention to the traffic. As part of that attentiveness, you could plainly see the dire predicament that was emerging in front of your very eyes.

Do you mind your own business?

Or, do you extend a helping hand?

Quite a conundrum, that’s for darned sure.

You could believe that those other drivers ought to be taking care of themselves. The person that opted to semi-park their car in a precarious spot will deserve whatever happens. The person making the right turn is also going to get whatever they deserve. If the driver making the right turn is careful, they should be able to notice the semi-parked car and come to a halt before hitting it, or perhaps swing wide to go around it.

All in all, this is none of your business and the world is what it is.

Wait for a second, suppose you did intervene. Perhaps you could save those fellow humans from getting into a car crash. There might be serious injuries involved that you could otherwise have aided in averting from happening. Just imagine how you would feel the next day, assuming you did nothing, and those people get hurt. Sure, it wasn’t your fault, but at the same time, you could have done something, nearly anything, for which the car collision might not have occurred.

What might you do, assuming you feel that doing something to help is warranted?

You could try turning your headlights on and off, flashing them at the car that was coming up to make that right turn. This might startle the driver and get them to slow down and become more cautious.

You could honk the horn of your car. The driver of the semi-parked car might hear it, and realize that it is a warning. Or the car making the right turn might hear the horn and opt to come to a full stop.

One supposes that you could try to move forward and start into the intersection, which might catch the eye of the driver making the right turn. This though also begins to put you into a dangerous spot. You are doing something ostensibly illegal, which might seem “reasonable” as an attempt to prevent a car accident from taking place.

Of course, your actions may be for not. The drivers of those other cars might not hear your horn or consider it irrelevant to their driving actions. The headlight flashing might not be seen by the driver of the car turning right. And so on. At least you would have tried to help. You would have the solace that you did the best that you could do in the given circumstances. You didn’t just watch blankly as things went sour. You took overt actions that any prudent and caring person might be expected to try and undertake.

Let’s add a twist.

Suppose your attempts to get those drivers to avoid being struck has an inadvertent and totally unexpected adverse consequence. Maybe the flashing of your headlights causes an entirely different driver to become distracted, and they plow into a car ahead of them. They later assert that their deadly crash only happened because they were preoccupied with your flashing headlights.

Whew, that’s a downer, and it shows that sometimes the best intentions can lamentably produce foul results.

Win some, lose some, you might say.

You have to take a chance and be a Good Samaritan when you can, could be one viewpoint.

There are laws in many jurisdictions that encourage the notion of being a Good Samaritan. The idea is that this will encourage people to help their fellow humans and do so without undue qualms about running afoul and suffering legal repercussions. Going even further, there are sometimes laws about having a duty to assist or a duty to rescue, whereby a person that can be of help is supposed to help if they can reasonably do so.

I’m betting that you’ve found yourself driving along and come upon instances of this general nature. You perchance noticed a car accident or car crash that was imminent, and you had to decide what to do about it. The pending car crash was not of peril to you. If you did nothing, you were perfectly fine. If you did something, you could also be perfectly fine. You had no direct stake in the outcome, other than the fact that you saw something and might have been able to avert the incident from occurring, though even this is a probability notion such that your actions might make a difference, or they might not.

One could suggest that not taking any action ought to lead to a guilty conscience. Maybe. For some, this could be a haunting situation for the rest of their lives. For others, they might contend that nothing they would have done could have produced any different outcome. Fate had already decided what was going to occur. Their watching the unveiling from a distance was a quirk of fate too.

Welcome to the woes of humanity.

Speaking of humankind, the driving of cars is gradually going to switch over to the use of Artificial Intelligence (AI) driving systems that will be driving cars for us. These are considered self-driving cars (for my extensive coverage of self-driving cars and AVs, see the link here).

An intriguing question arises: Will the advent of AI-based true self-driving cars include the AI spotting the emergence of imminent car crashes by nearby human-driven cars, and if so, what should the self-driving car do about it?

That’s something few are giving any thought toward, principally because we don’t yet have widespread self-driving cars. Once we do, this kind of matter is likely to come to the forefront. We might as well try to address the topic now, beforehand, and see what we can come up with. Welcome to the expanding realm of Ethical AI.

Before jumping into the details, I’d like to clarify what is meant when referring to self-driving cars.

Understanding The Levels Of Self-Driving Cars

As a clarification, true self-driving cars are ones that the AI drives the car entirely on its own and there isn’t any human assistance during the driving task.

These driverless vehicles are considered Level 4 and Level 5 (see my explanation at this link here), while a car that requires a human driver to co-share the driving effort is usually considered at Level 2 or Level 3. The cars that co-share the driving task are described as being semi-autonomous, and typically contain a variety of automated add-on’s that are referred to as ADAS (Advanced Driver-Assistance Systems).

There is not yet a true self-driving car at Level 5, which we don’t yet even know if this will be possible to achieve, and nor how long it will take to get there.

Meanwhile, the Level 4 efforts are gradually trying to get some traction by undergoing very narrow and selective public roadway trials, though there is controversy over whether this testing should be allowed per se (we are all life-or-death guinea pigs in an experiment taking place on our highways and byways, some contend, see my coverage at this link here).

Since semi-autonomous cars require a human driver, the adoption of those types of cars won’t be markedly different than driving conventional vehicles, so there’s not much new per se to cover about them on this topic (though, as you’ll see in a moment, the points next made are generally applicable).

For semi-autonomous cars, it is important that the public needs to be forewarned about a disturbing aspect that’s been arising lately, namely that despite those human drivers that keep posting videos of themselves falling asleep at the wheel of a Level 2 or Level 3 car, we all need to avoid being misled into believing that the driver can take away their attention from the driving task while driving a semi-autonomous car.

You are the responsible party for the driving actions of the vehicle, regardless of how much automation might be tossed into a Level 2 or Level 3.

Self-Driving Cars And Being A Good Samaritan

For Level 4 and Level 5 true self-driving vehicles, there won’t be a human driver involved in the driving task.

All occupants will be passengers.

The AI is doing the driving.

One aspect to immediately discuss entails the fact that the AI involved in today’s AI driving systems is not sentient. In other words, the AI is altogether a collective of computer-based programming and algorithms, and most assuredly not able to reason in the same manner that humans can.

Why this added emphasis about the AI not being sentient?

Because I want to underscore that when discussing the role of the AI driving system, I am not ascribing human qualities to the AI. Please be aware that there is an ongoing and dangerous tendency these days to anthropomorphize AI. In essence, people are assigning human-like sentience to today’s AI, despite the undeniable and inarguable fact that no such AI exists as yet.

With that clarification, you can envision that the AI driving system won’t natively somehow “know” about the facets of driving. Driving and all that it entails will need to be programmed as part of the hardware and software of the self-driving car.

Let’s dive into the myriad of aspects that come to play on this topic.

First, let’s dispel the myth that we are going to have exclusively self-driving cars on our roadways and somehow cast away all human-driven cars. Some pundits exhort that this is what we have to do. We must get rid of all human driving. By doing so, we will save lives from those foibles that humans have as drivers.

This dream world vision is rather outstretched. The number of self-driving cars is going to increase over time slowly, gradually, and incrementally and will not miraculously overtake all human-driven cars in one magical swoop. Humans will still be driving cars, for many decades, at least.

Keep in mind too that there are insistent car drivers that exclaim you will never take away their driving, ever never, until the day that you pry their cold dead hands from the steering wheel of their car.

All right, I think that sufficiently clarifies that there is going to be a mixture of self-driving cars and human-driven cars and that there is an opportunity for self-driving cars to indirectly help out human drivers.

How so?

Exactly akin to my earlier discussion about spotting a pending or imminent car crash, an AI driving system could detect those kinds of unfolding dilemmas. Pretend that instead of me sitting in my car as the driver at that left turn, a self-driving car was in that same position. The AI driving system might detect the possibility that those two cars across the street are heading toward a likely collision.

What would you want the AI driving system of the self-driving car to do about it?

Aha, this takes us back to the overarching theme, namely, should the AI mind its own business (I don’t mean that in an anthropomorphic sense, instead what programming of the AI system should we expect the developers to have put in place), or should it try to help out. You might also find of interest that this kind of open problem of a societal nature is similarly characterized by the famous or shall we say infamous Trolley Problem (see my analysis at this link here).

The answer by most of the existing self-driving car tryouts is that trying to figure out the imminent perils facing other cars is not within the pay grade of current AI driving systems. That’s a bit tongue-in-cheek, but it makes the point that the existing efforts of programming for AI driving systems are pretty much focused on what will happen to the ego vehicle, meaning the self-driving car itself.

There is plenty yet to be done to get all the programming done for that focus. We want self-driving cars that can go from point A to point B, safely so, and the attention of the designers and developers is honed to that goal. Having the AI look beyond the confines of what is relevant to the self-driving car would be considered an out-of-bounds type of problem.

In that way of thinking, this notion of spotting potential car crashes beforehand, which have nothing to do with the self-driving car at hand, would at best be on a so-called edge or corner case listing of things to do. This list would have various someday items on it. You can bet that the idea of programming the AI to ferret out dangers for other cars and pedestrians is quite low on the list, especially when doing the same for situations involving the self-driving car are obviously foremost and paramount.

Maybe, once we get the ego vehicle stuff relatively done, then the attention can go towards the considered “peripheral” action of helping out other drivers.

The problem too is the devilish details about what we would want the AI driving system to do anyway. Notice that we already covered the angst and agony of what a human driver is or ought to do, and did not reach a final consensus as to what must be undertaken. These leave open the question about what we want the AI driving systems to do.

One argument is that the AI driving system could be using up precious computer processing cycles by looking beyond its main scope. The computational effort to examine the driving scene and find instances of potential collisions that have nothing to do with the self-driving car could be said to distract the onboard computer resources.

Keep your eye on the ball, as they say.

We might though have some clever lawyers that argue a self-driving car could have detected an imminent peril and yet it did nothing about it. Returning to the example about my sitting at a left turn, a self-driving car in that same spot would have had an opportunity to detect the arising problem at the far corner of the intersection. This might be done to ensure that the self-driving car itself was not in any danger, meanwhile, the detection has taken place.

If the people at that far corner got hurt, and if the self-driving car could have detected the potential situation and took no action, this might be argued as a failing on the part of the self-driving car. Some believe we are going to hold self-driving cars responsible for their actions, as though they are the equivalent of a sentient being. This seems quite doubtful, and certainly not within the present-day realm.

Who you presumably can hold responsible would be those humans that developed the self-driving car and the humans that have fielded the self-driving car. Imagine a lawsuit against an automaker or fleet operator that failed to establish their self-driving cars with the capability to spot imminent car crashes, or that maybe had such a feature but did not use it to aid other travelers.

It is a somewhat novel idea and we’ll need to see how things play out. Realizing that the automakers and the fleet operators are likely to be bringing in the big bucks once self-driving cars are readily available, you can anticipate that these kinds of legal wrangling challenges are undoubtedly going to pop up.

Conclusion

There is a lot more to cover on this topic.

For example, consider this head-twisting barnburner.

There is an assumption that self-driving cars will likely make use of V2V (vehicle-to-vehicle) electronic communications. A self-driving car might spot debris on the roadway and emit an electronic message to warn nearby self-driving cars to avoid blockage. This is handy.

If a self-driving car was able to warn other self-driving cars about their getting into an imminent peril, but was not doing the same type of warnings for human-driven cars, this might cause quite a stink. It would be as though the AI driving systems are ganging up and helping each other, but not helping those human car drivers out there on the roadways.

You can just see the outroar. AI making life better for AI but shuns helping out humans.

Perhaps the AI might bemoan, well, guess you can’t win them all.