Autonomous vehicles – blessing, curse, or something in between?

Increasingly, it’s looking like autonomous vehicles are going to dominate our roads in just a few years.  What’s more, if you own a current-technology or earlier vehicle, you may not be allowed to drive it, thanks to advances in vehicle automation and autonomous control.  Mish Shedlock comments:

Capitalism is precisely why driverless is coming. Corporations are betting their money and resources. The government is not resisting. The trucking industry will save hundreds of millions of dollars. People who believe driverless is not coming are the ones who do not understand capitalism!

Fully autonomous vehicles are not some pie in the sky prediction by Al Gore. Real companies (hundreds of them) all working on driverless. A bet against them is a foolish bet against capitalism.

Comparing current carpooling with what’s going to happen is like comparing ancient stone huts to modern houses. Carpooling requires a number of people to get together, on the same route, for rides at the same time every day.

On-demand scheduling, point-to-point, is needed, and in the works. I rather doubt that fuel-based cars disappear by 2024, but widespread (not total) disappearance of privately owned vehicles by 2030 seems reasonable.

. . .

Some point to how few autonomous cars are on the roads. It all starts somewhere. In 1900, in New York City, there was not a car on the road. By 1920, there was not a horse in sight.

Others say they will never accept the technology. Perhaps they will when their insurance costs go through the roof.

There’s more at the link.  There’s also valuable information in these reports:

Both are worth reading.

I find these reports both interesting, and deeply troubling.  There are certainly many positive results that may come out of this, if the technological and legal problems involved can be solved.  Can autonomous vehicles be made ‘hacker-proof’?  Who’s legally responsible if autonomous vehicles are involved in an accident?  There are many questions like those that can’t be answered at present.

However, I’m also deeply concerned at the reduction in personal freedom and autonomy that this technological evolution represents.  Consider:

  • If I want to drive anywhere right now, I can get in my vehicle and go.  What if a government edict says I can’t?  Consider a scenario such as evacuating in the face of a hurricane or other national disaster.  If roads are blocked by too much traffic, it would be child’s play for some bureaucrat to digitally signal all vehicles in a large area, to allow only those within a given range of transponder ID’s to move at any one time.  An hour later, those ID’s could be blocked, and a new range allowed to move, and so on.  For that matter, the same technology could be employed to reduce rush-hour traffic jams every day.  From a bureaucrat’s perspective, this is a wonderful idea – controlling mass movements of people to prevent ‘disorder’ or ‘chaos’ . . . but what if they get it wrong?  What if their plans are overtaken by events such as natural disasters?  Besides, who gave them the authority to stop me going where I want to, when I want to?  You can bet we won’t be given a say in the matter!
  • What if I don’t want an autonomous vehicle?  What if I want to retain my existing, driver-controlled vehicle?  That may become impossible, partly because insurers will refuse to cover my old-fashioned, non-autonomous vehicle, and partly because manufacturers will no longer produce them, so that when mine wears out, I have no choice but to replace it with an autonomous model (if, that is, I can afford to – or am allowed to – replace it at all).  What’s more, cities may (and probably will) pass local laws to the effect that if you want to drive on their streets, you have to be in a vehicle that can be controlled by their traffic management systems, so as to prevent ‘disorder’ or ‘disruptions’ caused by ‘outdated technology’.  Present private vehicles may become automotive dinosaurs (not to mention their drivers!).
  • What if government decides to use vehicle autonomy as an extension of law enforcement?  In theory at least, any vehicle controlled by a traffic management system can be ordered to pull over to the side of the road and stop.  If there’s (say) a bank robbery, local cops could tell every vehicle within ten blocks to stop until they can check them all – whether or not they were involved.  If an agency wants to conduct a ‘safety check’ (whether or not that’s the real reason to stop vehicles), it can conceivably tell every car on the road to pull over at a designated point for inspection.  Drivers would no longer be in command – they’d be passengers, with no choice but to obey orders.  Some may argue that’s no real problem in a democratic country, but what if the checkpoint was in a totalitarian nation, looking for ‘enemies of the state’ (real or imagined)?  What if it were in a religiously fundamentalist nation, looking for those who profess other faiths, or who are classified as ‘heretics’ or ‘apostates’ by the powers that be?  Vehicle autonomy might become a tool of oppression rather than freedom, under such circumstances.

I’m not blind to the many advantages of autonomous vehicles.  As I get older, and become more infirm, it may be that an autonomous vehicle will allow me many more years of independence than I could have by relying on my own faculties.  This is a good thing.  I just hate the thought that autonomy will both increase our mobility, and decrease our individual freedom.  Is there no way to reconcile those conflicting values?



  1. I don't believe that it will happen as soon as many people seem to think. The current sensor technology is not adequate for this application. The example of the car not seeing the semi-truck turning across the road and decapitating the driver will be repeated many times over if there is a sizable number of self driving cars on the road.

    They also have great difficulty following badly or unmarked roads/lanes and do horribly in construction zones. The problem is that people are unreliable and bad drivers but computers are reliably good driver when things are clear but reliably bad drivers when things are not so clear.

  2. At most it will be localized, probably only in major cities. But the days of widespread, total automation is a very long way off if ever.


  3. I think it will happen, and in some areas where human cost is a big part of overhead it may happen relatively fast (why do you think both trucking companies and Uber are investigating this?).

    But I don't think it will be all over, or all at once, just as drivers now need to share the road with motorcyclists (who generally relish the riding experience), antique cars (I sometimes see Model A's on the road, even on the freeway occasionally), and even horses and buggies in some areas.

    I also think that people who use cars daily – in other words, people who don't live in the center of urban areas – are going to be very resistant to buying cars that don't give the option of manual control.

    Even if you live in the suburbs or the less-dense parts of major cities you routinely encounter situations where you need to apply human judgment on alternative routes or road conditions. Autonomous as an option sounds wonderful if you're making a long drive when you're tired, or if you're impaired (temporarily or not).

    But anyone who's ever been misled by a GPS, or just knows more about current traffic conditions (accident reported on the radio, ball game getting out, etc) has firsthand experience with the disadvantages.

  4. Taking away their private vehicles is one of the 2 things that would cause Americans to rebel.

  5. I see people predicting this, but I doubt it will happen as quickly as its supporters say.

    Despite all of the hype, many of the companies involved in autonomous cars do not have a background in life critical systems, unless manufacturers take safety more seriously, I expect to see a wave of accidents and fatalities that will create a pause in the industry.
    If large trucks go autonomous, this will get worse since they can cause so much more damage.

    Additionally, even in all vehicle were supposed to be autonomous and police controllable as you posit, there will be ways around the controls, analogous to the way that there are ways around the national internet filters in China, Saudi Arabia, etc – What one person can program, another person can break.
    One futuristic book I read a while ago with mostly autonomous cars still had manual cars for high security applications due to software risks (cash deliveries, VIP transport, police, etc).
    If autonomous cars do become widespread, there will be pressure to adopt them but unless our country changes substantially, it won't get to the level that you describe for a long time. I can see where certain states and cities will push their adoption, like they have emissions regulations, but if you live outside those areas things will be different.

  6. "If I am even allowed to" That is the very reason why you have Islamic terrorist killing at will in London and you live in the states and not in Durban. My statement on autonomous cars? OH HELL NO! They are in and of themselves grounds for civil war. We should have burnt it all to the ground over mandated health care. I'll burn in hell before some government SOB tells me what I can drive. Cars and guns are the two things that government had better leave alone lest they start a civil war.

  7. I agree with Anonymous at 2:06 PM. There is no better way to get pig hunting season to open, with no bag limit.

  8. I'll believe driverless cars when they can do driverless trains. At least those have relatively simple traffic patterns to work with.

  9. Peter, a good point about evacuation. One school of thought is that the driverless cars will optimize i.e. Reduce the physical car inventory, as the majority of cars spend most of their lives parked. An evacuation would require all the current capacity at the peak. Does this mean that some people may not be removed from danger?

  10. Look at aircraft to see where automated piloting/driving is heading. There have been reasonable efforts to fully automate cockpits, with some success, for 20+ years. Planes are FAR easier to automate vs automobiles. Still not 'there' with planes. Mass transit will be first, and already exists on a very small scale with airport shuttles and such. I could see large city mass transit getting it done in my lifetime. Then maybe some success integrating the suburbs. When they can integrate my truck backing a boat down a ramp or backing up to hitch the trailer, well, then maybe we're getting somewhere. Now the 'self parking' cars are a novelty, and attempts at high speed auto control are already resulting in tales of severe wrecks. Not 'there' yet…not really even close. IMO the automakers are putting far too much $$$ into this tech and hurting their stock prices, i.e. Ford.

    Just 2 cents from a petrochemical instrumentation and controls engineer…

  11. @Timbo, yes, it would. Though the inventory of cars would no doubt be larger than enthusiasts dream of simply because it doesn't matter what percentage of time a car is unused now, they still have to be able to meet peak demand every single weekday. Cars that spend a much higher percentage of time being used will also be out for maintenance and repairs much more. Doing away with privately owned, manually controlled cars is a Utopian pipe dream for a couple of decades, at least, but one fervently wished for by authoritarians who know what's best for everyone.

  12. because I believe in capitalism, I believe that we will not get driverless cars for many, many years, if ever. Corporations invested in (and .gov backed) electric vehicles and they are a very small part of the overall car sector. Americans like our own independence, and there are many, many jobs for which carpooling, or automated systems won't work for. Capitalism works on what folks buy. I don't see too many folks buying an automated car. Those corporations better have a way to get the money back, otherwise, they could lose their shirts.

  13. I can see big cities doing it, a bit like Germany with it's emissions zones and fees (the more pollution your vehicle emits, the more restrictions on where you can go in places like Berlin and Frankfurt.) Out in the sticks? Eh, not really. Even if the tech reaches the working point. And all it will take is one f-, ahem, fouled up hurricane evacuation, or people not being able to get TO people in need (tornado response, flood response) for backlash to hit.


  14. From a bureaucrat's perspective, this is a wonderful idea – controlling mass movements of people to prevent 'disorder' or 'chaos' . . . but what if they get it wrong?

    What do you mean if?

    Germany will do it first, maybe a couple of other Yurps afterward. I'd guess local, low speed, delivery vehicles might be among the first in the US, plus dense-urban taxicabs; UPS and FedEx probably are already helping to finance some of the deivery vehicle research. As for OTR trucks, rather than being completely autonomous, I'd figure a train of 2 or 3 "slaves" behind a human-driven "master."

    Humans are just too unpredictable and software too stupid for autonomy to work unaided. I wouldn't be surprised to see a great deal more hardware/software assist provided to humans long before full autonomy.

    As to the fear of government control, expect a burgeoning underground army of HW/SW mechs to develop "drive control switch apparatus" to defeat bureaucrats.

  15. There was a large group of people in New Orleans that depended on government controlled transportation. As Katrina bore down on the city, those responsible for running that transportation ran. After the storm passed, as the slow motion disaster of the levees breaking happened, the government did not marshal the transportation under their control to evacuate those people and they were left to die.

    Autonomous vehicles, like going cashless, require a fully functioning electrical and data network grid. In addition, autonomous vehicles require a near perfect road system, without wash outs, flooding, damaged sensors and signals, etc.

    And, of course, this presumes that never again in human history will one group be denied the same privileges as another by government.

    "Similarly the American who has been humbled by poverty, or by his insignificance in the business order, or by his racial status, or by any other circumstance that might demean him in his own eyes, gains a sense of authority when he slides behind the wheel of an automobile and it leaps forward at his bidding, ready to take him wherever he may personally please. "

    –'The Big Change: America Transforms Itself 1900-1950' (1952), Frederick Allen Lewis

  16. Autonomous vehicles are coming, but until AI improves it's not likely to be widespread. Ride sharing will only ever be effective in urban and suburban settings. It's simply not practical in rural areas. As for now, and in the near future, as much as I get upset with my vehicles when they break down or need routine maintenance, they are still far more inexpensive and convenient than ride sharing or public transportation.

  17. I had the idea of for a murder mystery where the victim was placed in a self driving car then given a slow poison. The car would then be programmed around the intersection of three or four states. When the car and body is finally discovered, no one can determine where the crime actually occurred.

  18. The whole idea of autonomous cars will last right up until some hacker loads one up with explosives and runs it in to the House or Senate office building. What a stupid idea. Sorry Citizen, you are not cleared for that destination…Sorry Citizen, you have been ordered to the work camp. I will be shooting long before.

  19. The people designing autonomous cars or delivery vehicles for the likes of UPS or FEDEX have never been down the gravel/dirt county road or up the private lane to my house. In mud season? In winter? Maneuvering around washouts? I live on a decent rural road, the sketchy ones are much worse. Good luck with that. I would be curious to see how these autonomous vehicles deal with driving in heavy snow, ice, dust storms, even on a normal highway. One US highway in my area gets water over it during floods which require a state pilot truck to guide the traffic through a foot+ of standing water for a mile or so. Last week while on a trip we had to drive through several inches of storm water on city streets with leaves, sticks and other debris floating around. How's an autonomous car deal with that? It's all well and good on dry pavement and grid square streets with nice fresh reflective pavement markings and a good GPS signal. What happens when those things aren't there to hold their robotic hand?

  20. Tou don't even need bad roads. Read elsewhere the following scenario:

    The Big Game is over. Thirty thousand people pour out of the stadium, pile into their automated cars, and say "Home, James."

    Which car pulls out of its parking space first?

    Several hundred questions follow that one. At least.

  21. Cars will be programmed to stop so as to not hit /injure a pedestrian. It will take only 2 nanoseconds until the "teens" know that they can bum rush a car and jack it. You will be immobilized and surrounded.

  22. This technology, like ALL technology is neutral, neither good or bad.
    It is the agenda of those who control the technology and those who
    legislate that determines the good or bad that technology inflicts on
    society. Those who legislate specifically seek that power because they
    are almost invariably greedy, selfish and evil. They seek positions of
    power because they lust for power…..for control over others. It is
    THIS group that will generally determine how a technology is used and
    who if any suffers or is deprived of rights. Till we find a way to keep
    these petty tyrants seeking to become GRANDE tyrants out of office (or to
    kill them outright when they display their true colors) there is no solution
    to how technology is used as a weapon against us and our freedom.

  23. Having recently worked on one of the big railroads here in the U.S., I find that automation of train movement is still not fully functional. And that's with over thirty years of work in the development of, and twenty years of deploying such systems, on tracks mind you. I would not worry about true automatic pilots for cars for at least thirty years, if not more. Then trying to get people to adapt to such systems will take even longer. Let's not forget the laws will need to be changed at federal, state, and local levels, which can take even longer. And, you can always use a horse and buggy if things really get out of control.

  24. Autonomous cars.

    Sure. You bet.

    They'll work – sort of, within limits, under certain condiitons, within these parameters, etc. – in an environment suitably sterilized for them. Maybe in an extremely well structured "urban core" they'll kind of work, just as some cities provide public bicycles for rent in their "urban cores." In that same environment they may "kind of work" for some sort of delivery function; 30 years ago some large businesses and government agencies used automated – "automated," very decidedly not "autonomous" – mail carts, which, using very basic infrared technology, followed a painted trail on the floor, periodically stopping and "bonging" outside an office door to announce "the mail is here." Some offices were allowed 1-2 minutes to come retrieve their mail, others were important enough that the cart waited until someone pressed the "continue' button. That'll certainly work well at the corner of Third and Main; whatever locking compartments it may have will simply be sales opportunities for crowbar merchants.

    "Autonomous cars" is a reality-starved egghead's attempt to digitize what is entirely an analogue environment; Mother Nature is not nearly so accommodating as a well-lighted, perfectly straight hallway in a temperature and humidity controlled building with security officers at the front door. Or, even a perfcectly developed "urban core" with very precisely defined curbs, intersections and the like, much less any road surface with imperfections, or even texture variations or random adjacent obstacles like unmowed grass that's 12 inches taller than it was last month. They'll continue to work perfectly, I'm sure, in every George Lucas movie. Out here in Reality Land, NFW.

    Remember how, before skateboards "went pro" and became big money items, kids used to scrounge wheels from shopping carts to make them? Autonomous cars will become rolling store houses for the scavenging of various parts long before they develop sufficient utility to become widespread for their intended purpose.

  25. I don't think it'll happen any time soon, I don't think the state and local governments will ALLOW it to happen, and for only one reason: They'd lose all revenue from traffic and parking tickets. After all, if your car is under some master controller, it can't speed, it can't change lanes without signalling, it can't park in a no-parking zone. Plenty of cities and towns make MAJOR money off of such violations (oh, they couch it as public safety, but it's really about revenue). For instance, New York City makes about half a BILLION a year in parking tickets. Just parking, not other traffic offenses like speeding.

    Governments don't like to lose revenue.

    Mark D

  26. I wonder how the automatic car idea will be accepted by communities such as the Amish: plain folks living simply? Or by the aged widow(er) who has just barely mastered the touchscreen telephone technology. And don't get me started on the rural roads in my county of 11,000 citizens. Daunting situations for the adoption of auto-car technology.


  27. re Anonymous' "bum rush":

    Roger Zelazny and Fred Saberhagen did a book with a side note about automated trucks on dedicated superhighways. It included a bunch of teenagers across the country who'd started up outlaw clubs based on a peculiar kind of performance art.

    You wait by the side of the road. At a particular, carefully calculated moment, you cross the road. At a certain speed, possibly varying it at certain points. The oncoming truck doesn't stop–you are just a smidge too far away to justify that–but it slows down, then speeds up again when collision is no longer probable.

    Several miles down the road, another kid will do something similar. And another further along. And so on.

    Eventually you get "wave action" in the traffic. If you do it right, you can cause major traffic snarls, analogous to hammering in water pipes. Kind of like the variable speed limits they do now to smooth traffic on inner-city expressways, but used for evil.

    Hack the system without any access to it. Just to show you can.

  28. Self driving trucks will start showing up fairly soon, but they will not displace that many drivers.

    For one thing there is a significant shortage of drivers already, for another, long-distance highway driving is easy to automate. maneuvering around in narrow city streats, in parking lots, loading docks, etc is very hard to automate.

    What I expect to see happen is humans will still be in every big rig, but as passangers for most of the time, only taking control when conditions are bad (including at each endpoint)

    This will allow a given truck to be rolling a much higher percentage of the time, which translates into big bucks, but still with a person on board to provide security and handle the corner cases.

    shared car fleets have a good chance of impacting taxi companies in big cities, but much less chance of replacing commuter cars.

    on the other hand, personally owned self-driving cars can make big inroads in commuter traffic (allowing those who can afford them to do other things while on the road.

    I really don't think that insurance companies will drive traditional cars off the road, they aren't going to raise insurance costs from their current levels, they may very well give self-driving cars a discount, but that doesn't make traditional cars cost them more (and there is still competition in auto insurance, so as long as one company is willing to make less money per driver to get more money overall by getting more drivers, cost of policies will stay stable with respect to the costs of accidents)

    As for the point that in 1900 there were no cars in New York, and by 1920 there were no horses, once you look outside of the dense urban areas, you will see that horses stuck around as a common mode of transportation until after WWII

    David Lang (Commercial Class A Hazmat driver)

  29. I'll believe when somebody comes out with an affordable method to transport me and my recently shot, field dressed and dirty deer / wild pig / elk / moose that I just shot back to my house from a mile down that two track road out in the sticks. Furthermore, how is the farmer going to get his cow/pig/sheep etc to the butcher?

    It's one thing to transport people… but it's going to be another thing to tell the rednecks to give up their pick up trucks!

  30. I'm an engineer at a company working on cameras and ladar systems for autonomous cars, and while it's correct that there's a lot of companies pushing a lot of money into this tech, I don't think we will see widespread adoption of driverless cars in the next 10 years. There's still a ton of work to be done and really hard problems to solve.

    By the way, the idea that disagreeing with the hypothesis of rapid onset of driverless cars is somehow going against the consensus of capitalism ignores the fact that we see plenty of cases where the consensus is wrong and/or has the timing incorrect. The massive investments in fiber in the tech boom 20 years ago was a classic example. Good idea, but quite a bit too early, and a lot of companies and people lost their shirts.

  31. @Clinton J
    the same can be said for many groups of people, including the mom transporting the soccer team of just a family of very young kids (with all their toys, supplies, etc)

    anyone who needs to have much equipment/stuff readily available will not go with a shared driverless car. The types of people who have heavy/messy stuff to move will not be welcome by the shared car companies.

  32. Autonomous driving on the freeway under normal conditions is something that can be done pretty reliably today, and if this allows the person to do other things, than this will be a desirable feature (similar to how self-parallel parking is already).

    for Trucking, this is a large percentage of the miles driven, and if it can be done without the drivers hours-of-service limits kicking in so that a single driver can keep the rig rolling 20+ hours a day with a single driver (instead of requiring teams of drivers and being lucky to have it rolling 15 hours/day) there will be a lot of companies willing to invest in this.

    David Lang (commenting as unknown so that I can get the follow-up comments e-mailed to me)

  33. Actually, automated driving on freeways is probably the easiest task.

    And the one that I suspect most people would be most willing to adopt. There have been many times on long trips where I needed to stop because I kept nodding off, or just needed a short mental break. And – especially when I was younger – many times I probably should have pulled off, but didn't. Or I arrived at a destination city mentally exhausted even before I had to deal with heavy traffic.

    Being able to have the car take over handling the long-and-grueling part of driving while letting the humans handle the challenging parts is a much easier sell than convincing people to give up control.

  34. I am enormously skeptical about autonomous cars. The easy problems are easy (and have been solved, allowing flashy demos). The hard problems (including security) are hard. Some are really hard, and will kill people.

    At that point the issue will shift from technology and transportation policy to liability. I'm not worried about losing my insurance on my driver-required car, because the insurance companies will be VERY sure about their liability exposure before they pull that plug.

    But your point remains that companies are in a big hurry to do this. Ford just fired their CEO and replaced him with the head of their autonomous car unit:

    Me, I think that the liability from autonomous car crashes will put at least one automobile manufacturer out of business. It might be a truck manufacturer, but it's hard to see that someone won't get taken down from this.

    Think Donald Rumsfeld's "unknown unknowns".

  35. Autonomous cars? We'll see. I can tell you a few things for certain sure:

    Senior citizens will cheerfully accept an autonomous, self-driving car rather than give up their driver's license, and that gives the government one certain inroad to getting people to use the autonomous car – clamp down on seniors that are marginally able to handle today's traffic situations.

    I'd take one in a heartbeat. I hate driving in traffic, and at the outset autonomous cars will be given preference over anything with a lunatic behind the wheel. Just tell me how many homicidal or suicidal or just plain crazy-stupid drivers you see during rush hour in a major city. Add to that the times I'm down at the bar and would like to have another 3 or 4 shots, but can't because I'm driving, and there you go.

    When it gets to the point where everyone is either forced to have an autonomous car, you'll see motorcycle riders protesting. I'll probably be one of them.

  36. "I don't think the state and local governments will ALLOW it to happen, and for only one reason: They'd lose all revenue from traffic and parking tickets. Mark D."

    The communities of NJ would go bankrupt (NOT AN EXAGGERATION!), as they acquire a significant portion of their operating funds through this means.

    They would have to figure out how to tax those autonomous vehicles to recover the lost revenue.

  37. They would have to figure out how to tax those autonomous vehicles to recover the lost revenue.

    Their attempts to regulate CO2 and penalize those who produce it prove that they've figured out how to tax air (it's just a matter of implementing it in such a way that the people don't lynch them.)

    Taxing autonomous vehicles will be child's play. "Charge per mile traveled" most likely, since they'd have all of the tools needed to track it. They could even vary the rates based on the type of road.

  38. @LindaG, in places like New York, that's likely to be electricity (short ranges, lots of stop-and-go), a bit harder for them to tax more

  39. Since they will probably only roll them out in big cities, that is true. However, as they found out when Trump was elected, there are a lot of places that are not big cities.
    Who knows. They may do something stupid like the Cash for Clunkers program when they start rolling them out, just to sucker, um, pull people in.

  40. for shared cars to be successful, it's going to need to be private companies doing this (ala Zip Car) so that they can refuse serviceto people who abuse things (leave the cars a mess, etc). Such companies may end up merging or setting up cross-company shares (think the equivalent of cell phone roaming)

    But it's not going to be something setup and run by governments, so I'm really not worried about them requiring it.

    People aren't going to jump on the 'new greatness' if it doesn't suit them, the slow uptake in electric cars is a perfect example. For some people, they are the best thing ever, but those folks usually have multiple vehicles and use the electric for the commute, frequently with their job/office providing recharging.

Leave a comment

Your email address will not be published. Required fields are marked *