The headlines proclaiming the emergence of self-driving cars are usually soaked in high praise for the advent of these technological marvels.
This heralding of such wonderment in innovation is certainly understandable.
Getting a car to be driven entirely by automation and altogether absent of human hand at the wheel is an enormous feat, let there be no doubt or question on that account. Other than dreamy portrayals in sci-fi novels and movies, the notion that a car could safely and sufficiently make its own way on our public roads is a stretch goal, some assert it is a moonshot-like aspiration, and one that has alluded inventors and systems developers since the automobile itself was first rolled out for human use.
If you are looking for reasons to want self-driving cars, a boatload exists.
Pundits are anticipating that we might find ourselves entering into an age of mobility-for-all.
The idea behind this mantra is that the availability of self-driving cars will dramatically increase access to transportation. Self-driving cars will be continually on-the-go, 24 hours per day and seven days per week, roaming our streets and highways so that they are nearest to you when a lift is requested. Unlike human-driven cars that require a human operator at the driving controls, a self-driving car does not need any rest breaks per se and can just keep on trucking, as they say.
The hope too is that the cost of using self-driving cars will be substantially less than the cost of using human-driven automobiles.
By subtracting out the labor involved in driving a car, the assumption is that the use of these AI-driven cars will be a lot less costly per mile. People today that are unable to afford the use of a car, such as existing ridesharing services, will find themselves readily able to pay the much lower cost of a ridesharing self-driving car (not everyone agrees on this cost aspect, including qualms that self-driving cars might be overly expensive and only for the rich or famous, see my analysis at this link here).
With the huge upside of self-driving cars, the celebratory welcoming of self-driving cars is indubitably sensible and deserving of outsized headlines.
That being said, some have expressed and continue to express deep concerns about the arrival of self-driving cars all told.
We all seemingly realize that with any new innovation there are going to be positives and also likely negatives. Any realistic real-world invention that touches mightily upon society has got to have rough edges and downsides. There is no free lunch in this world, and that certainly includes self-driving cars.
It is easy to dismiss out-of-hand some of the objections raised against the introduction of self-driving cars.
For example, sometimes the objectors proffer nothing other than vacuous statements that land as hollow accusations without any substantive foundation. Anyone merely touting that self-driving cars are “bad” and yet does not provide a basis for such a boisterous claim, well, this kind of empty argument does little and regrettably gives the more reasoned naysayers a bad name.
Another set of oppositional arguments drift toward sci-fi possibilities.
This includes the postulated future of AI becoming sentient and then opting to use self-driving cars against humanity (see my discussion of these theories at the link here). AI is not sentient at this time. We don’t know that AI of the kind that we are devising will ever be sentient. The AI driving systems that are being put in place today are absolutely and abundantly not sentient, and there is no miracle breakthrough seemingly required for reaching sentience in order to adequately drive a car (for more about this hotly debated point, see the link here).
These aforementioned ways of trying to warn us about self-driving cars are not particularly successful in their meaning and nor merits.
Despite those shortcomings, we ought to not let that cloud our thinking on this quite important topic.
In short, there are assuredly bona fide reasons and a poignant rationale for those that seriously and sincerely have doubts about the benefits of self-driving cars versus the “costs” of self-driving cars (I’ve put the word costs into quotes since this is not just a dollars and cents consideration, as there are downsides or disadvantages that are not immediately relatable in monetary terms alone).
Perhaps I’ve now whetted your appetite to find out about the more rigorous arguments in opposition to self-driving cars.
They are coming up next, I promise, but first, some added and crucial contextual milieu.
These are arguments that can each of their own accord can be separately debated and in some ways separately discounted (a classic divide-and-conquer form of unraveling an argument).
Tackling each of the arguments individually or on a standalone basis is an easier way to undercut their total argumentative power or sway. As such, it might be instructive to try and weigh together the entire set of arguments.
Out of this final weigh-up, the scales can ergo be more notably claimed to tilt toward not unleashing self-driving cars, so some proclaim.
Take a reflective moment about that somewhat startling assertion.
Essentially, the claim is being made that there are a core set of fundamentals that inform as to not permitting self-driving cars to proceed at all. The declaration is straightforward and directly grasped: Do not let self-driving cars get underway now, and ostensibly not in any foreseeable future. Don’t let self-driving cars be dribbled out. Don’t let them gain any grassroots.
That undoubtedly seems a bit of an extreme and nearly counterintuitive posture.
This relatively stark viewpoint that self-driving cars ought to not be made available is quite a far cry from stating that before self-driving cars are put into use that certain steps or mitigations ought to be established. As you will see, the counterargument is that it makes no sense to start something that opens essentially a Pandora’s box.
Once you’ve let the genie out of the bottle, it isn’t going back in, and turning back is nearly impossible. In that frame of mind, consider this collection of reasons to not let self-driving cars proceed as not just a red flag, but instead, a permanent stop sign that shall never be passed (of course, “never” is a long time, and maybe we’ll all change our minds in some distant future).
Before leaping into the arguments, one aspect to clarify involves the definitional aspects of what it means to refer to a self-driving car.
Let’s unpack the self-driving car definition first, and then get into the guts of the arguments for fundamentally preventing them from being placed into use.
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 Why Don’t Proceed
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.
The crux of the basis for not wanting self-driving cars is due to the AI doing the driving (as I early exhorted, this is not because the AI possesses sentience or is omniscient, so please do not anthropomorphize today’s AI capabilities).
The primary source of concern is that a human won’t be doing the driving and that instead, a form of automata will be doing so (a form of automation that is devised by humans, will be overseen by humans, will be managed by humans, and so on).
I mention this to also help clarify that a self-driving car is still a car.
In other words, the automobile itself is not particularly different than a conventional car. Sure, a self-driving car has a suite of specialized sensors such as video cameras, radar, LIDAR, ultrasonic units, and the like, but that isn’t especially unlike a regular car. Indeed, the first wave of self-driving cars is almost certainly going to be conventional cars that have been retrofitted with the sensors and the computing processing units needed to enable the on-board AI driving system to function.
Furthermore, even once we begin to see new kinds of cars that are devised from the ground-up as “pure” self-driving cars, those will still nonetheless be cars. They will have four wheels, they will have an engine, they will pretty much be the same as conventional cars are today. Sure, the interior of self-driving cars is likely to be reimagined, allowing for riders to swivel while seated in a self-driving car and also recline for catching a few winks while the AI is undertaking the driving chore. This is still the essence of a car.
The point is that there is no need to look at the car (as an automobile) for being the culprit or gotcha in not wanting to have self-driving cars.
Look at the AI driving system, that’s where the thorny problems ensue, as will be discussed now.
Strap on your seatbelt and get ready for a heck of a ride.
Arguing About The Lives Lost Or Saved
For those that favor the advent of self-driving cars, there is a simple and commonly declared reason for wanting self-driving cars, namely the saving of lives.
Here’s the logic, and then we’ll see how the side that oppositionally disfavors self-driving cars offers an intriguing retort.
Currently, in the United States alone, there are annually about 40,000 car crash-related human-fatalities and around 2.5 million injuries due to car crashes. A sizable portion of those car crashes is due to human drivers that are drunk while driving. Another sizable portion is attributed to human drivers that are distracted from the driving task. An AI driving system is not going to be a drunk driver. An AI driving system is not going to be a distracted driver.
The assumption is that the use of self-driving cars will significantly reduce the number of car crashes, thus saving lives and foregoing the concomitant injuries. No more human drivers that are drunk while driving. No more human drivers that are watching cat videos while driving. And so on.
That does make sense.
But there are those in the favoring camp that take this a bridge too far and say that we will have zero deaths and zero injuries because (allegedly) self-driving cars will never get into car crashes of any kind.
For various reasons that I’ve carefully articulated, this zeroing mantra has a zero chance of occurring (see my explanation here). The notion of zero crashes is certainly alluring, though it also presents utterly false expectations. There will irrefutably be some amount of self-driving car crashes, beyond zero, though we don’t know how many and how frequently it might occur.
Okay, so if you’ll agree that there will be some amount of self-driving car involved car crashes and related human fatalities and injuries, which is the likely and realistic scenario, this is where the disfavoring arises.
Those that fundamentally oppose self-driving cars would argue that any deaths of humans at the hands of an AI driving system is one too many. AI should never be allowed to kill a human, as per the classic Asimov’s rules of robotics (see my coverage at this link here). Until an AI driving system can be guaranteed on an ironclad basis that it will never ever get into a car crash, and thus would fatality harm a human, such AI driving systems should not be allowed on our public roadways.
That is a showstopper.
A child is standing between two parked cars and cannot be seen from the active roadway. A car is coming along at the posted speed limit. The child suddenly darts into the street. No matter how responsive the car is, the child in this scenario is going to get run over and sadly become a fatality, even if the car immediately applies the brakes.
Suppose this was a self-driving car. That is undeniably a self-driving car that has killed someone.
Case seemingly closed.
Of course, a human driver could equally be in the same quandary and have produced the same result.
The assertion by those that disfavor self-driving cars would be that though the result might be the same, the apparent fact that an AI driving system was at the wheel is intrinsically wrong and showcases that the AI driving system is undertaking the sacred duty that should be reserved exclusively for humans and human decision making. This gets the other camp in arms since they point out that the net benefit of the number of lives lost will be likely tremendously reduced. In other words, the AI driving systems will save lives. It might not prevent all lives from being lost, and yet any lives saved are precious and ought to outweigh the fact that self-driving cars can nonetheless still be attributable to some lives lost.
For those that believe that even one life lost due to an AI-based self-driving car is too much, albeit that perhaps many lives will have been saved, this is a key cornerstone for disfavoring or outright opposing self-driving cars.
There is a quagmire of ethical considerations amidst all of this (see my columns for further details).
We’ll move to the next argument for now.
Arguing About The Extinction of Human Driving
At first, there is going to be a mixture of human-driven conventional cars amidst the advent of self-driving cars (there are about 250 million conventional cars in the U.S. today). Gradually, the number of self-driving cars would be increased and thus the proportion of human-driven cars versus self-driving cars would shift accordingly.
Some have urged that we ought to curtail all human driving on our public roadways and enforce that only self-driving cars could use the publicly shared highways, streets, and byways. The logic is similar to the foregoing discussion about the dangers of drunk drivers, distracted drivers, etc. Remove all human drivers and all of those untoward driving actions would presumably no longer exist.
Assume that indeed human driving becomes summarily banned (well, maybe with some exceptions, such as driving allowed on private property or for old-fashioned human-driven car races).
If that ban occurred, in many respects this would be the extinction of human driving.
No one would be driving a car (setting aside the small exceptions). The bulk of society would no longer know how to drive, either forgetting how to drive or never having learned to drive since there was no need to do so. In any case, driving a car on public roadways would generally be considered an illegal act, subject to criminal prosecution and incarceration, one presumes.
The disfavoring camp emphasizes that this is a denial of what some would characterize as a foundational right for mankind, namely the act of driving a car (even though, admittedly, it is currently classified as a privilege, and not an inherent right). Some have suggested that the day you’ll take away their driving is the day that you pry their cold dead hands from the steering wheel.
As a counterargument, it is said that we don’t know that such a ban will occur.
Meanwhile, the counter to the counterargument is that it is a likely slippery slope and an obvious insidious outcome of the advent of self-driving cars.
Arguing About The Dependency On Machines
Envision that nearly all use of cars involves self-driving cars.
Imagine that very few if any cars remain that are able to be driven by humans. The days of conventional cars that contain a steering wheel and pedals have faded into history. For all practical purposes, using a car entails having an AI driving system do the driving task.
This is similar to the prior discussion about banning human driving, but we don’t have to go that far. Just assume that there aren’t many cars with driving controls, thus, even if a human wanted to go for a drive, there aren’t any cars around to do so. You don’t particularly need to ban humans from driving if by default there isn’t an available means to do the driving.
This can be claimed as supporting the disfavoring case, here’s why.
Want to go for a drive on a whim and go wherever you want to go?
Since you can’t do the driving, you’ll need to tell the AI driving system where you want to go.
Change your mind during a driving journey and want to take the next exit to see something interesting that caught your eye. You can’t opt to do so. That being said, yes, you can tell, or ostensibly “ask” the AI to take that next exit for you.
Humans will become dependent upon machines, the AI driving systems, and self-driving cars, and will no longer be in charge of their own driving pursuits and interests.
Tying this point to the prior one, this also bodes for the problem that if human driving is extinct, and we somehow suddenly need to be able to drive because the machines aren’t or won’t (this is somewhat inching onto the sci-fi quicksand, but not entirely so), humans won’t be ready to do so.
The decay in driving skills or the absence of driving skills cannot necessarily overnight be overcome.
Arguing About The Traceability
Today, you get into your car and drive to the store, then over to the neighborhood park, and then head back home. Nobody other than you is aware of your travels.
With AI driving systems, the odds are that they are going to be keeping track of every ride and every inch that is being driven. This data will then be uploaded to the cloud of the fleet operator. You are going to be tracked to a degree that boggles the mind.
Do we want that kind of traceability of every driving journey and personal choices of where we go, when we go there, etc.?
That seems like a quite compelling argument for the disfavoring camp.
A counterargument is that if you use ridesharing today, this same kind of tracking occurs.
Aha, the counter to the counterargument is that the preponderance of driving in the United States today is not ridesharing, which is a drop in the bucket. By-and-large, our driving currently is done via our individual and relatively untraced personally owned cars.
It is likely that self-driving cars, regardless of ridesharing use or personal use, are going to always be tracking their every activity and driving effort.
Arguing About Collective Control
You decide to drive up to the mountains where you have a vacation home. Along the way, you hear on the radio that a forest fire is underway in that area. Though you know there is risk involved, you drive there anyway, for reasons of wanting to protect your property and reach some beloved animals being kept there.
In an era of self-driving cars, here’s what might happen.
The government decides that it is too dangerous for people to be going into the mountains that have forest fires. Contact is made with the major self-driving car operators that have the largest fleets of self-driving cars. They are told to command the self-driving cars to stay out of those now-restricted areas. Electronic missives are transmitted to the self-driving cars and the AI driving systems are informed to not drive in those restricted areas.
Your choice of going up to the mountains has been summarily taken away from you.
Whereas in a conventional car, you could make that choice, the AI driving system as instructed by the fleet operator is not going to proceed there.
This particular example allows for easy debate about whether the government is making a prudent choice and perhaps saving lives by enforcing such an edict. Suppose though that the government in other circumstances started restricting where you could go and did so on not quite such an apparent lifesaving basis. There might be no end in sight for where you could no longer go, and for which there won’t be much you can do about it (assuming that there aren’t human-driven cars available and that you cannot override the AI driving system instructions).
The advent of AI-based true self-driving cars can potentially produce a form of centralized control over driving that is nearly unimaginable today. It might seem laughable by today’s standards, but if you look closely at what some major state-actors are considering doing with self-driving cars, this kind of emergent centralized control is decidedly not a farfetched notion (see my column for examples).
Arguing About Societal Downgrading
In interests of saving ink, I’ll lump together a bunch of additional disfavoring arguments into this one category of societal downgrading or downfall aspects.
Here they are in rapid succession:
· Learning to drive a car is considered a societal rite of passage, but that won’t happen anymore in an era of self-driving cars. That is ostensibly a societal downgrade or downfall that could be construed as an indirect adverse consequence of self-driving cars.
· No more sense of independence comes with being able to drive a car and go where your heart desires.
· We might undermine some of our cognitive skills by no longer learning to drive and carrying out the act of driving.
· The same might be said of the accompanying fine motor skills of handling the driving controls.
· Some believe we might become lazy, using self-driving cars when we ought to be walking or riding a bike.
· You can readily sleep while inside a self-driving car, thus, perhaps we’ll become more sloth-like.
· You can easily go for a drive yourself via a self-driving car. No driver’s license is needed and no one else is required to drive the car for you. This might splinter us into becoming more alone.
· Children can be driven around by a self-driving car. No need to have a parent do this. Some wonder if this might undermine the family bonding that normally entails the act of using a car for family activities.
· People might eschew using mass transit and flock exclusively to self-driving cars, thus undermining and likely ultimately decimating existing mass transit and the future of such transportation.
There are lots of these kinds of societal ramifications that no one knows whether they will arise or not.
Another facet to keep in mind is the possibility of bad actors that are somehow able to overtake the cloud of a fleet operator or that can manage to implant unsavory code into the AI driving systems. This is yet another disfavoring point that can be added to the mounting heap of such foundational objections.
Conventional cars of today are currently relatively untouched by this kind of viral attack. That being said, even conventional cars are gradually adding interconnectivity and tons of on-board processing, such that they too are going to have these kinds of potential vulnerabilities.
Overall, now that you’ve gotten a ton of bricks dropped on top of that outright joyous glee over self-driving cars, does that crush those ambitions, and should we bring things to a Luddite-like halt?
My take is that the future will be what we make of it. These disfavoring assertions are worthy of our rapt attention, now and later on, and we ought to not undervalue or avoid the difficult questions that have to be properly addressed.
The glass is half-full, but it is also half-empty, and we need to be cognizant accordingly.