Word To The Wise: Do Not Try Asking An AI Self-Driving Car To Park In The Red
You are giving a friend a lift in your car and as you approach their desired drop-off there is no apparent place to idle or park to allow them a convenient exodus from your vehicle. So, you decide to come to a brief halt at a curbside that is painted red.
Is that a right or wrong action?
More importantly, perhaps, is this a legal act or an illegal act?
In your mind, even though you acknowledge that a red zone is not to be used for this purpose (well, depending upon the state and local jurisdiction), you nonetheless feel fine about doing so since it will be only a momentary stoppage, plus you are still actively at the wheel of the car and the engine is running.
No big deal, you convince yourself, and furthermore, there are tons of people that do the same thing all the time.
In California, the official vehicle code law says this about curbs painted red: “Red indicates no stopping, standing, or parking, whether the vehicle is attended or unattended, except that a bus may stop in a red zone marked or signposted as a bus loading zone.”
This seems to clearly answer the question concerning whether momentarily coming to a stop at the red curb is legal or illegal, whereby it is ostensibly illegal unless somehow you want to claim that your car is a bus (good luck on that one!). Imagine telling a judge that since you were dropping off a passenger, your car counts as being a bus and therefore you were perfectly legit in your actions. Probably this is why judges contend they have heard it all.
In any case, the law in California is seemingly and abundantly clear cut and for which you cannot stop, you cannot come to a halt or be at a standing, and nor can you park in a red zone. I realize that you are likely immediately imploring that you see people do this all the time. As such, the law is presumably stupid, ill-advised, and the will of the people seems to have decided it is not worth observing. This is reminiscent of your parents that probably used to say that just because someone else jumps off a cliff or leaps off the end of a pier it does not mean that you should do so too.
People that tend to use those red curbs are apt to claim innocence.
Your honor, I knew that you couldn’t leave your car parked there while unattended, but I had absolutely no idea that it was also illegal to be inside the car and parked there. How might this plea go over? If the judge got up on the right side of the bed that morning, perhaps you’ll be granted leniency. On the other hand, you might hear the famous refrain that ignorance of the law is not an excuse. Perhaps the judge will throw the book at you for your lack of driving rules awareness.
Yet another trick is to argue that the red curb was not a red curb, or that at least the red paint is so faded and chipped that it was not obviously apparent that it was a red curb. Maybe you might try too to exhort that you are colorblind and therefore the red looks to you like it isn’t there are all. Akin to a magician, this is the proverbial rabbit-in-the-hat ploy of trying to convince the judge that the curb wasn’t red or that nothing in this world is really red (a lofty philosophical argument).
Here’s a particularly vexing curb-related case in California that especially catches the eye.
A woman parked her car in a spot wherein the curb was definitely not red. The authorities agree that she parked in a location consisting of a non-painted curb. While her car was parked there, the city came along and painted the curb red. Pictures of the car now parked in the newly painted red zone showcase that the red paint on the curb skirts around the point of her tires that were just touching the curb. There seems little dispute that the red was painted after she had earlier parked her car there.
Do you think she should get a ticket?
This seems egregiously unfounded, though the authorities allegedly stated that she got the ticket for being parked in that one spot for too lengthy of a time period and not due to being parked in the red. She meanwhile asserts that the ticket states that she was ticketed for parking in the red. It’s all a mess.
For those that fervently believe in the red curb, they always make sure to never park, stand, or stop at a red curb. Never ever. Furthermore, those ardently faithful to the red zone are apt to indicate that scofflaws and those violating miscreants ought to get whatever is coming to them (maybe a hanging or the electric chair?).
I know one driver that recently looked in disgust at a car that was momentarily stopped at a red curb, which seemed to be only doing so to disgorge a passenger that needed extra help and could not otherwise walk a greater distance. The observing driver of the red curb stationed car spoke aloud that this was an illegal action and he frantically looked around to spot a cop that might rush to issue a ticket.
Upon not seeing any nearby police, the next comment was that perhaps it might be useful to position his car to block the vehicle that was sitting at the red curb. The logic seemed to be that this would elongate the time of the illegal procedure and give a greater chance of a ticket giver to come along. But, in the end, it became just another example to this person of how far gone we are in terms of obeying the law, allowing lawlessness to exist.
In this viewpoint, the red curb facet is a microcosm of a bigger societal problem. Letting red curb lawbreakers get away with this illegal act is tantamount to having a snowball that is rolling down a hill and becoming larger and larger as it does so. If people can commit the crime of misusing a red curb then they are bound to ultimately commit more heinous digressions.
Okay, so some people look at a red curb as simply a suggestion about what to do, while others see it as a highly visible and readily apparent indication about the exactitude of the law and a need for us all to stridently abide by our laws (or perish otherwise).
What is your viewpoint about red curbs?
Always obey? Sometimes obey? Ignore the red and do as you wish? Find a means to catch and punish all those that transgress the red? There are a dizzying variety of perspectives about those red curbs.
In thinking about the future, one aspect that comes to mind is that we will eventually and inevitably have true self-driving cars on our roadways.
This brings up an interesting question: Will the advent of AI-based true self-driving cars imply that there will never be a violation of a red curb again, namely due to the AI aiming to always be fully legally compliant with the driving laws?
Let’s unpack the matter and see.
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 a 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 a 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 Those Red Curbs
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.
Some believe that the advent of AI-based driving systems means that we will have strictly legal-oriented drivers from now on. No more running of stop signs. No more zipping through red lights. No more speeding above the posted speed limit.
No more parking, standing, or stopping at a red curb.
End of story.
Maybe, maybe not.
First, keep in mind that the AI is going to presumably do whatever it has been programmed to do. In that manner, if indeed the AI developers programmed the AI to not break the law, in theory, that is what will happen. Of course, like any programming, there are potentially embedded errors or omissions and we do not know for an absolute certainty that this legally minded logic will be fully embedded and observed.
Notice that I am not suggesting that the AI is somehow sentient and will change its mind. Contrary to what you might have heard or seen on the news, there isn’t any AI yet today that is sentient, and we aren’t close to getting there. There isn’t any AI that has common-sense reasoning and nor AI that can be said to approach the full gamut of human intelligence.
Thus, do not worry that the AI has a mind of its own and will do whatever it wants. There is no mind in the AI per se, at least not yet.
The AI can nonetheless go astray, due to errors in the code or other problems that can arise. These anomalies are not mindful ones. They are aspects that arise due to the complexity inherent in millions upon millions of lines of coding.
There is another path for the AI to opt toward illegal driving activities such as parking or idling at a red curb.
Via the use of Machine Learning (ML) and Deep Learning (DL), much of what the AI driving system is doing can at times be based on the computational patterns that it has gleaned over time. By providing lots of driving examples to the AI system, it can attempt to identify patterns of driving and then presumably seek to drive as based on those found patterns.
Here’s the rub.
Suppose that the driving data consists of the efforts of human drivers and their driving actions. In a large trove of such driving-related data, there are bound to be indications of some that have opted to park or stand at a red curb. Envision that the AI finds this pattern and therefore assumes that this is an allowed driving action.
Those that are developing and testing the self-driving car might not be able to discern that the AI has a deeply embedded calculation that includes being allowed to stand or park at a red curb. The odds are that the everyday use of a self-driving car during this experimental stage of being on our roadways might not by happenstance cause the AI to park at a red curb. It could be that this golden nugget, as it were, sits deep inside the arcane AI artificial neural network and has yet to be invoked.
At some point in time, the AI might come across that embedded aspect and then choose to stand or park at a red curb. If so, would there be anyone around to tattle on the self-driving car? Would the fleet deploying the self-driving car realize this has happened?
Also, suppose that in one jurisdiction you can legally make use of a red curb for standing, as long as the vehicle is considered as being attended. Does the AI driving system count as being the attending driver? Presumably, yes, since the AI is doing the driving. On the other hand, the AI is not the same as a human, in that a human does more than simply the act of driving, and perhaps the gist of having the car be considered attended is because we expect the driver to be all-around aware of what is going on.
That raises a larger AI & Ethics related set of questions and a great deal of attention is now taking place on these timely and crucial topics (see my analysis at this link here).
You might be tempted to suggest that the AI developers ought to be able to inspect their AI code and ascertain whether it has all of the properly coded legal driving rules. This is a problematic concern with the use of Machine Learning and Deep Learning.
For ML/DL, usually, the computational morass that embeds the foundations of the driving task is not readily explainable. Think of this as a bunch of numbers, piled on top of each other, out of which there is no easy way to grasp what it portends in any symbolic or logical semblance. Known commonly as the interpretability or explainability problem, there is extensive research going on to try and craft XAI, explainable AI (see my coverage at this link here).
What About Passengers Wants
Slightly shifting gears, consider the role of human passengers whilst inside a self-driving car.
A self-driving car is providing you with a lift to your office. Turns out there is a red painted curb directly in front of the main office entryway. Further down the street is an unpainted curb. Today there is a horrific storm and the rain is coming down in buckets. You don’t have an umbrella with you.
When the self-driving car nears your destination, you speak to the AI.
As a side note, the expectation is that self-driving cars will be making use of Natural Language Processing (NLP). Similar to an Alex or Siri, the AI of the self-driving car will talk with you. You might tell the AI that you want to have the self-driving car swing through a Starbucks on the way to work so that you can get a cup of java. Some of the automakers and self-driving tech firms have been making a false assumption that there won’t be much if any interaction between the AI and the passengers. This falsehood is based on the narrow thinking that all you will do is get into a self-driving car and it will take you to your designated destination. This belies the reality of how we ride in cars and there is oftentimes a significant amount of interaction between the passenger and the human driver about driving-related matters.
In any case, back to our story about going to work on a rainy day. You tell the AI that it should drop you off directly in front of your office, which you wish to do because it is rainy and you don’t want to get wet by being dropped off further down the street.
Will the AI be able to detect that the curb is red?
If it detects that the curb is red, will it determine that this is not a legally proper place to stand or park?
And, if it does all those things, what should it do about the human-uttered passenger instruction?
You might wish to argue that the AI ought to do whatever the human passenger has told the AI to do. We should presumably expect that the AI will be obedient to humans. Imagine a world gone amuck if the AI won’t do what humans tell it to do.
Turn out there is a slippery slope involved.
Consider the use case of a human that tells the AI to go faster than the speed limit because the person is late for work and wants to try and get to the office as soon as possible. Should the AI proceed to drive faster than the speed limit?
Worse still, imagine the use case of a human that tells the AI to run over someone. Should the AI proceed to run over a human being? Your reaction in this use case is that there is never a basis for running over a human and therefore this is obviously a verboten command. You might want to read my coverage of the case involving a shooter on a bridge and for which a human driver ran down the shooter and saved many lives accordingly (see my discussion at the link here).
The point of those examples is that the AI presumably should not abide by the requests or commands of human passengers without some kind of consideration about the bonafide nature of those indications. Unfortunately, trying to make decisions about those human commands is a lot harder than you think it is, and for which (as mentioned earlier) there isn’t AI that has any common-sense reasoning or sentience to engage in a mindful dialogue on such topics.
For now, the answer by the automakers and the self-driving tech firms is that if a passenger wants to do something beyond the norm, the AI or the passenger will invoke a kind of OnStar-like feature that connects the rider to a fleet operator. The agent answering the connection will seemingly engage in a discussion and then potentially use a remote means to give guidance to the AI driving system.
That seems like a nifty solution, but it is not especially eloquent and at times altogether impractical.
The timing of the need of the passenger might not allow for the making of the connection and dialogue over what is to be done. Also, the remote agent might not have the capability to direct the AI driving system in the manner that the passenger hopes (plus, there are concerns about security and other facets of using a wide array of remote commands on an AI driving system). There also might not be a viable communications connection possible between the vehicle and the remote agent. And so on.
It might seem surprising that something as simple as coming up to a red curb could become an entire entanglement about the nature of AI and the role of self-driving cars.
We have a natural tendency to take for granted the capacity of humans to drive a car. On the surface, all that driving seems to require is the ability to press on the gas pedal or the brakes, and merely steer the wheel of the car. The reality is that there is a lot more that goes on in our minds and that ultimately drives the driving actions that we take while driving.
The next time you come up to a red curb, consider carefully what you will do.
Maybe even ask yourself, what would the AI do?