News report of toddler pushing a stroller that was saved by a driver from wandering into traffic is … [+]
getty
A driver did the right thing the other day and saved the life of a meandering toddler.
Per recent news reports, a driver in Virginia noticed that a little girl was getting dangerously close to the roadway. The toddler was pushing a child’s toy stroller and had a stuffed animal inside the petite carriage. She and her comforting companion (the stuffed animal) appeared to be in their own playful world, enjoying a casual frolic outdoors and not mindful of the cars, buses, and trucks whizzing nearby.
There didn’t seem to be any parental supervision in the proximity of the unfolding and altogether ominous calamity.
The cute youngster might have been so distracted by her stroller and her beloved stuffed animal, and equally unaware of the perils of being in traffic that she might have veered directly into the pathway of lumbering vehicles. We don’t know if other drivers would have noticed her. The worst outcome could have occurred, namely, an unthinkable collision could have arisen.
Let’s not go there.
Instead, we can relish that a driver decided to stop his vehicle and take quick action. He was already wearing a neon vest since he was driving a bus. Upon stopping the bus to try and block traffic, he then scooted out and waved at oncoming drivers to maneuver over and out of the way. Meanwhile, he scurried over to the girl and guided her to safety.
He and another good Samaritan proceeded to walk with the girl on a door-to-door basis in the neighborhood to see if they could reunite the wayward girl with her family. They were able to do so.
A happy ending ensued in that the toddler and her stuff animal with stroller were all preserved and averted a serious and somber outcome. One can only hope that the parents will take this as a blissful lesson and ensure that the toddler does not get into such a predicament again.
Would you have stopped your vehicle to aid the youngster?
I’m assuming that nearly all of us would absolutely do so.
Perhaps that is overly optimistic. I realize that you can argue about whether or not taking such action was entirely warranted per se.
I mention this because the child wasn’t yet fully into the street. She was nearing that dire position. It would be easy to have assumed that she would not go directly into the roadway. Perhaps her instincts would prevent her from doing so. She might have also been taught to stay out of the street. All of those are distinct possibilities.
The question entails when do you decide that enough is enough. Is it sufficient that the child appeared to be aiming toward the roadway? Or do you believe that until an actual entry into the street occurs is there nothing out of the ordinary to be done?
You can’t seemingly stop at all points in time and places that you suspect someone is going to potentially mistakenly go into the street. You would never make any driving progress. That doesn’t seem like a practical everyday solution.
Whoa, the retort goes, this was just a toddler, perhaps around no older than 3 years old (per the estimate of the heroic driver). If you see a toddler that is wandering on their own, diverging toward the street, and assuming that you don’t observe anything that will prevent a doomsday outcome, you have to take action. Come on, the lament goes, show some compassion and common sense. This is ostensibly a rather unique situation that necessitates that type of boldness.
Let’s recast the scenario and pretend that a parent had been nearby.
In that case, you might have frantically waved at the parent or honked your horn to alert the presumably distracted or otherwise preoccupied adult. That would be the first step of action. You would not have anticipated that you would need to get out of your vehicle entirely. Assuming that the parent still seemed to be lacking in care, you might certainly bring your vehicle to a complete stop.
And at that point, as a final act of heroism, possibly get out to somehow more directly get the parent to watch out for the child, only as a final step of exasperation and concern.
Maybe the parent would be irritated and upset that you intervened. The toddler was just fine, the parent might insist. This is none of your darned business. Of course, had you not stopped and not proffered assistance, and if the toddler had gotten into the street and injured, you would forever be alarmed that you didn’t take overt action. Perhaps the parent might even accuse you of not having made an attempt to alert them about the impending catastrophe.
But returning to the actual situation, we know that there wasn’t a parent or any adult within reach.
In that circumstance, another question that arises is whether you would notice that the child was indeed alone.
Think about that for a moment. You are driving your car, minding your own business. You are watching the other cars. You are watching the trucks and buses. You are watching bike riders. There is a lot to keep your attention riveted to the roadway.
You glance at the sidewalk and take notice of obvious instances of endangerment. For example, a dog that is on the loose and running directly for the street. That would probably catch your eye. You might spy a jogger that is wearing a headset and seems oblivious to the traffic conditions. If that person, even if an adult, suddenly turns as though they are going to enter into the street, I’d bet that would cause you consternation and you would definitely notice the adverse action.
There are zillions of instances of pedestrians on the sidewalk that might not especially put you on alert. You cannot be watching and mentally tagging every nuance of the sidewalk activity. Doing so would seem to distract you from the traffic itself. No sense in ramming into another car simply because you are preoccupied with whatever is happening up on the sidewalk.
In short, we have some drivers that might not have noticed the little girl at all.
Yes, that’s totally feasible.
Envision that we somehow did an experiment and stopped those cars further down the street. We might ask the drivers; did you see that toddler on the sidewalk? Some percentage will say that they didn’t, and some of those might insist that there wasn’t a child anywhere in the vicinity (claiming, defensively, that you are making it up, trying to fool them).
That is one segment of drivers.
Another segment might have noticed the little girl but had not noticed the absence of an adult being present. Again, envisioning a kind of experiment, we would likely find that those drivers assumed that a parent must be nearby. In their minds, a child on the sidewalk is almost always accompanied by an adult. The adult must have been there. Some of those drivers will reluctantly concede that they didn’t see an adult, while others might exhort that there was an adult there and you are lying that an adult wasn’t present.
This is all the stuff of human behavior.
Let’s try another variant.
There might be drivers that noticed the little girl, plus observed explicitly that no parent was around, but these drivers did not deduce that the girl was heading toward the street. It would be easy to merely have subliminally acknowledged the toddler and yet not placed the matter onto any kind of driving priority list in your noggin. The toddler wasn’t acting outrageously or doing anything that would create abundant attention.
Those drivers would probably, later on, reflect upon the situation and kick themselves for not having stopped. They would replay the moment in their minds, revisiting the aspect that the child was alone and nearing the street. Most of those drivers would probably feel sick to their stomach about not having done anything. Some would proffer an excuse that they were preoccupied with some work-related matters or personal troubles that allowed them to let the toddler situation seem distant and inconsequential at that moment.
Well, welcome to just another day of driving. You never know what drivers are going to notice. You don’t know what is on their minds. Some drivers are highly attentive. There are drivers that are distracted. You’ve got all kinds of drivers, good and bad, considerate or abrasive, and so on.
The list is nearly endless.
Furthermore, the act of driving is one that plays out over time. This means that a driver might not be distracted at one instant of time, and then seconds later be deep in thought and become distracted. You would be hard-pressed to say that a driver was always one way or another.
That being said, let’s be honest and agree that there are drivers that are more prone for example to being distracted drivers. Someone that routinely glances at their smartphone while at the wheel of a car is more likely to be predominantly and continually distracted. A driver that almost never looks at their smartphone would be less so a distracted driver (at least with respect to the smartphone aspects), though they too might become distracted when the smartphone rings and suddenly draws their awareness.
Are there ways to cope with the human foibles of drivers?
Yes, there are.
Perhaps the most all-encompassing solution is to obviate the need for human drivers all told. In other words, excise the role of human driving and you presumably and by definition no longer have to cope with the failings of human drivers since there aren’t any human beings sitting in the driver’s seat.
Seems pretty straightforward and indisputable.
This is a perfect segue into talking about those emerging AI-based true self-driving cars.
Note that there isn’t a human driver involved in a true self-driving car. Keep in mind that true self-driving cars are driven via an AI driving system. There isn’t a need for a human driver at the wheel, and nor is there a provision for a human to drive the vehicle. For my extensive and ongoing coverage of Autonomous Vehicles (AVs) and especially self-driving cars, see the link here.
Here’s an intriguing question that is worth pondering: What can the heroic tale of the driver that saved the toddler and the stuffed animal inform us about the advent of AI-based true self-driving cars?
I’d like to first further clarify what is meant when I refer to true 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 The Toddler With The Stroller
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 is 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, it is important to realize that not all AI self-driving cars are the same. Each automaker and self-driving tech firm is taking its own approach to devising self-driving cars. As such, it is difficult to make sweeping statements about what AI driving systems will do or not do.
Furthermore, whenever stating that an AI driving system doesn’t do some particular thing, this can, later on, be overtaken by developers that in fact program the computer to do that very thing. Step by step, AI driving systems are being gradually improved and extended. An existing limitation today might no longer exist in a future iteration or version of the system.
I trust that provides a sufficient litany of caveats to underlie what I am about to relate.
We can use the toddler situation as fodder for discussing what today’s AI driving systems are seemingly capable of. This is a useful means of gleaning the weaknesses and potential strengths of using AI driving systems, albeit with my earlier caveats still being kept in your mind.
Would a true self-driving car detect a toddler on the sidewalk that is pushing a stroller of which contains a stuffed animal?
That’s an interesting question.
Self-driving cars are outfitted with a variety of sensors. The choice of sensors to be included is up to the automaker and self-driving tech firm. Generally, you can expect that the sensor suite would include video cameras, radar, LIDAR, ultrasonic units, thermal imaging devices, and the like.
I’ll go out on a limb and declare that it is likely the sensors would indeed detect an object on the sidewalk that was the toddler. There would be a blob of a child-like shape, along with a blob that is the stroller. The chances of determining that the stroller is a stroller are less likely. The odds of figuring out that the child is a child is somewhat higher, all else being equal.
My point is that the data collected by the sensors is merely data. The AI driving system has to try and analyze the data to find recognizable patterns. Of those patterns, the AI driving system can then categorize or classify what is found by the sensors. One would reasonably assume that most of the self-driving cars are doing sufficient pattern matching via Machine Learning (ML) and Deep Learning (DL) to determine that the blob that has two little legs and two little arms and a head is a child.
This is actually a lot harder to computationally calculate than you might assume.
A misclassification could easily be performed. There are potential false positives and false negatives that can be computed.
Also, a classification usually has a probability associated with the matter. For example, the blob might be categorized as a child, based on a probability of 85% chance that this is a correct estimation. Note that means that the classifier is not absolutely affirming the categorization and instead basing the result on probabilities and uncertainties.
Okay, we will focus on the toddler, and for now, we’ll ignore the stroller. The stroller is likely considered an unknown blob that seems to be attached to or somehow associated with the categorized blob of the child. The nature of this unknown blob is unknown, but it appears to accompany the child wherever it goes.
I’ll toss a bit of a twist at you. A human would presumably instantly realize that the stroller is a stroller. Also, it is a reasonable assumption that a human adult driver would know what a stroller is and what can be done with it.
Allow me a tangent to explain.
The toddler could presumably push the stroller toward the street and let go of the stroller as she does so. In that case, the child might remain on the sidewalk. Meanwhile, the child has sent the stroller as a kind of unguided missile into the traffic. A human would know that this could happen. Few of the existing AI driving systems have that kind of “reasoning” capability.
You see, one of the biggest challenges and qualms about being able to achieve true self-driving cars is that we have a huge unresolved form of human reasoning that commonly is referred to as common sense. Numerous efforts have been made to devise AI that can reason in a common-sense fashion. These unfortunately are not anywhere close to attaining human levels. See my discussion and analysis at this link here.
Back to the mainstay of the elicitation.
The AI driving system has likely calculated that the child is a child, but has no particular semblance about the blob that is the stroller. Via the sensors, we’ll suppose that the pace and direction of the child are being observed as moving toward the street. The prediction portion of the AI driving system is likely to calculate that the child is on a dire course toward the street and a potential intersecting path of the self-driving car.
Here’s the part that might get your goat, as it were.
By and large, most of the AI driving systems are focused on whatever might be an endangerment to the self-driving car. Does the child pose a danger to the self-driving car in terms of possibly colliding with the self-driving car?
If the answer is yes, the AI driving system will begin to perform whatever evasive maneuvers have been programmed for avoiding striking objects, including and in specific a person as a type of object in this instance.
If the answer is no, namely that the child would not be an endangerment to the self-driving car, the AI driving system is likely to be programmed to simply carry on with the driving task. In essence, suppose that the AI driving system computed that at the speed of the autonomous vehicle and its direction, it was not feasible for the child to end up in the street and be in front of the self-driving car.
That means that there isn’t any danger associated with a potential collision. Thus, keep driving. There is nothing else here to be done. It is like the famous Star Wars aspect of Yoda saying look away since there is nothing here to be seen.
I assume you realize the gaping problem with that kind of assumption. Sure, the self-driving car might scoot on past without any troubles. At the same time, the child might veer into the street just after the self-driving car has passed by, and the car behind the self-driving car might strike the wayward child.
In one sense, you could argue that some human drivers would do exactly the same thing in that they would opt to drive past the child and do nothing about the matter unless they believed that their car was going to collide with the toddler.
I don’t think we would hold such a person up as a model citizen per se.
Do we want our self-driving cars to drive in that manner?
That’s the million-dollar or more like a billion-dollar question, based on the millions and billions of dollars so far expended toward the attainment of self-driving cars. For my detailed exploration of this concern, see the link here.
Conclusion
There are a lot more twists and turns on this topic.
Due to space limitations, I’ll just offer a few closing remarks for now.
Suppose that the AI driving system did opt to stop the autonomous vehicle. This might be done to try and avert striking the child. It might also be done to try and block upcoming traffic that is behind the self-driving car. All in all, the self-driving car could be a kind of shield.
That’s what the bus driver did.
Of course, the bus driver also got out of his vehicle, flagged at other traffic to get those other drivers to slow down and move over, plus he then guided the child away from the street.
There isn’t any driver in a self-driving car to do any of those actions. Until the day that we end up with robots that can drive, which is something I’ve covered in my columns, see the link here, for example, the AI driving system is not a walking-talking contraption.
One supposes that the AI driving system could alert authorities by sending out an electronic message. The police might then arrive to aid in the situation. That seems sensible, though consider the time crunch involved. By the time the authorities arrive, it could already be too late, sadly so.
Another possibility is that if a passenger is inside the self-driving car, the passenger might realize what is taking place and rush out of the autonomous vehicle to take care of things. But we don’t know that a passenger will always be inside a self-driving car, and indeed the expectation is that self-driving cars will be roaming around a lot of the time with no passengers, see my explanation at the link here.
Having a human at the wheel is pretty handy in a situation of this nature. The human can be both a driver and act in other ways too, such as a hero that rescues an unaware and unsupervised toddler. Not all drivers would do this, of course, and you could therefore argue that the AI self-driving car fits into that percentage of circumstances.
We also have the foibles of human drivers that need to be taken into account. Humans are both a benefit and a drawback when residing in the driver’s seat. Getting rid of human driving and replacing the driving action with an AI driving system will provide some benefits but will also engender some drawbacks.
Let’s end this on a pleasing and altogether happy note.
We now know that there is a toddler that is alive today due to the heroism of a driver that opted to do the right thing. This particular toddler might be our future, as it were, growing up and someday be perhaps the inventor, discoverer, or AI developer that devises a level of AI that can embody common-sense reasoning and perform amazing acts that only humans today can do.
She and her stuffed animal could make the world a better place. Thank goodness that she survived that moment of dire consequences as a toddler and lived to achieve such greatness.
I assuredly like that outcome.
