Cars vs. Phones: Why Robocars Will Not Save Us

What with all the attention being paid right now to the struggling global economy, various wars, the 2012 electoral food fight, and Ben Flajnik’s underhandedness, the dawn of the robot car era may have slipped under your radar. Well, pay attention! Robocars (as I call them) have made a whole slew of breakthroughs in the past few years. They’re surprisingly close to appearing on a freeway near you. Anyone who wants to get up to speed need only check out Tom Vanderbilt’s rather glowing article in Wired. Vanderbilt is the author of Traffic, and one of the more thoughtful thinkers about the transportation system. If he’s impressed, we all should be.

It probably goes without saying, but compared to any of the glowing robocar pangyrics, my take is a bit more cynical.

Every time a text rings, someone may become an angel.

As Vanderbilt suggests, and as I’ve argued here at in the past, robocars point to a key paradox at the heart of 21st century society. There’s an unavoidable conflict between cars (the definitive 20th century technology) and smartphones (the definitive 21st century technology, so far). The more people are distracted while driving, the more accidents happen. That’s why it’s so eerie and problematic when companies like Mercedes start putting wifi in their cars and Facebook on their dashboards. That’s NOT a good idea!

The “cars vs. phones paradox” is perhaps one reason why a company like Google is investing so much time and money into robocar technology. For Google, not only is the robocar a great techno-utopian demonstration project, but it begins to solve a fundamental problem for internet companies. And Google isn’t alone. Car manufactures from Germany to China are focusing on robocars. Their very future may depend on it.

There are two huge problems, though, with the robocar concept in the real actual world. And other than a few brief asides, these problems aren’t really being talked about.

The first is the huge quantitative and logistical complexity of the US automobile system. There are over 200 million cars on the road in the US right now. We’ve gotten to the point, thanks to huge government investments spanning a century, where cars are a default mode of movement for almost every American, no matter what income level. Cars equal mobility and movement, and allow people to access jobs, food, schools and just about everything else too. They’re not a privilege in America today, they’re a right. That’s why gas prices, snowplowing, parking spots, and traffic jams are such hotbutton political issues. Many a political career have been built atop the automobile.

The massive number of cars involved – over 254 million registered – means that any new technology is going to take a long time to get implemented across the country. Think of all the old clunkers on the road. Think of all the cars that don’t receive proper maintenance. They aren’t going away. This doesn’t include the uninsured unregistered cars, which according to some estimates, are 16% of the cars on the road and rising dramatically every year. There are cars in my working class neighborhood that are literally held together with duct tape, or where the windows are plastic sheets.

All these facts mean that any change to the US auto system that’s as fundamental as the robocar revolution will not happen quickly or neatly or evenly. Robocars will create a two-tiered system of auto transportation that will be obvious and stark. There will be robocars for the rich – Mercedes-Benzes and Audis and GoogleToyotas – and used Hondas for everyone else.

Is this just the beginning of a car segregation system?

Not only will this pose problems for the promise of robocars, dramatically reducing many of their efficiency and safety gains, it will create a deep divide within the social geography of our transportation system. It’ll be a lot like at airports, where you have different lanes for business travelers and the masses. It’ll be like the difference between limos and bus stops. In an era when more and more Americans are out of work, overworked, or facing decades of declining wages, spending increasingly precious tax money on transportation investments for robocars is going to become politically and socially problematic. It’s something that we need to think about.

At this point, you’re probably thinking that this is yet more neo-luddite social justice nitpickery. And, sure it is.

You’re probably thinking that, of course the transportation system is huge and complex, and populated with a billion (no exaggeration necessary) constantly moving independent variables, e.g. potholes and skateboarders, each different and unique and unevenly distributed throughout society. But, you might say, this is solvable given vastly improved supersmart technology and effective and sound government policy. (That second one is a big if, by the way.)

The second main problem with robocars, though, is potentially a dealbreaker. That’s the question of who is liable when something goes wrong.

There are currently 32,000 fatalities on the US roads every year, and over ten times as many injuries and accidents. Just as an exercise, multiply EACH of those incidents by a million dollars, and you start to get a sense of the potential liability of the US automobile system. (It’s mindboggling number in the hundreds of billions of dollars.)

I’m not a lawyer, but under our current political and social system, most of those potential liability dollars don’t exist. Almost all the fatalities and injuries on our roadways are classified as “accidents”. They’re seen as inevitable by-products of a road system that’s impossibly complicated, ubiquitous, and disbursed across hundreds of millions of different individual actors, all of whom can’t be responsible all of the time. We write off all these accidents as “the cost of doing business.” It’s a social choice we’ve made, a trade-off for the mobility and “freedom” offered by the car.

This whole system depends on the profound lack of responsibility at the center of the picture. This whole system depends on placing a inevitably distractible and semi-non-responsible driver at the center of the picture. (The exact responsibility of the driver is an area of much heated dispute, but our social and legal system has basically given drivers a free pass almost all of the time.)

What if every time your computer crashed, your car crashed?

But what happens when you take that driver away, and let them start to doze off or read the newspaper or surf the web? Who’s responsible then when something goes wrong? Sure, in a robocar society, accidents will be greatly reduced.

But they’ll still happen! (And, this is what none of the dozen or so robocar articles I’ve read don’t bother to mention.) Something will go wrong. With 200 million cars and millions of other changing variables, something will always go wrong. Nobody who’s ever seen the “blue screen of death” or the “spinning beach ball of stasis” can possibly disagree. (What happens when a robocar system gets too old, and doesn’t get maintained? What happens when a small part breaks, and nobody notices?)

Well, the obvious answer is: the car company / robocar software company. If GoogleToyota is driving your car, what happens when something goes wrong and your car runs over someone’s grandma?


How many cars are recalled every year? How many more would there be if automaker liability was increased 100-fold?

A good analogy for this problem might be the US aviation system. Airplane crashes really disturb people, and for good reason. So, as a result, airplanes come with two pilots. And because even they doze off or use their computers while flying the plane, we have a ‘backup system’ costing billions of dollars in place, staffed by highly trained people whose only (very stressful) job is to coordinate 28,000 daily US flights.

Now multiply this problem by three or four orders of magnitude (1.1 billion daily auto trips), throw in a countless dogs and kids and deer and bicycle messengers and old ladies and out of control shopping carts and huge potholes and drunk drivers. How many air traffic controllers would you need?

The thing is, we already have an “air traffic controller” whose sole job it is to pay attention all the time to the world around them, and to keep their foot on the brake pedal in order to make sure that nobody is going to get killed or horribly injured by a 2500 pound steel machine moving at 50 miles per hour. That person is the driver. Do we really want to let them start taking naps?

There is a lot of promise behind the idea of robocars. But to my mind, most of the good effects of them seem to come from reducing the need for cars, not increasing it. For example, picture a world without parking spots. That world is possible with robocars. They can become a shared commodity, and not something that we all need to have all of the time.

Well, if that’s the great benefit of robocars, why bother? We already have a “robocar” technology that works really well, that allows you to surf the internet, sleep, or read while you travel. It’s called the bus. It’s called the sidewalk. It’s called a bicycle or a subway. All these things already exist!

For a time, my dad had a Cadillac Fleetwood Brougham with an “automatic trunk closing” feature. You would lower the lid of the trunk to a certain height, about one inch from the top of the lid, and then the car would “take over” and close the rest of the trunk for you. If that’s not the stupidest innovation in the history of technology, then you’re reading the Skymall catalog. Who had trouble closing their trunk in the first place? When the robo-trunk mechanism inevitably broke, my father had to spend a thousand dollars to fix it.

Maybe the robocar is a solution in search of a problem. If the problem is that our current road system is unsafe, unhealthy, bad for the planet, and incompatible with new technological norms, there are two roads before us. The first is to develop highly complicated new technologies that will “solve” the problem. While I’ve wary of the complexity of the situation, I’m not saying we can’t do this. Computers are amazing, the US is the wealthiest country in the world, with the best engineers and software designers you’ll find on planet Earth.

But for the cost of a single backup camera, you could buy someone a really nice new bicycle. For the cost of a single robo-interchange, you could implement a system of safe and comfortable bike lanes all through an entire city.

This is just one example of the road less traveled. If we go down this road, we start reducing our dependence on technology. Instead of robocars, let’s build cities that privilege people. Instead of thinking about ever more complex ways to depend on the automobile, let’s start thinking beyond it. To me, that’s a visionary future. It might not be as shiny, and it might not be up there with the Jetsons or Norman Bel Geddes’ Futurama, but it’s far more equitable, practical, and affordable than the robocar alternative.


The robo-car of the future. (PS. I did NOT have them pose for this picture. This is how young people ride the bus.) is a non-profit and is volunteer run. We rely on your support to keep the servers running. If you value what you read, please consider becoming a member.

18 Responses to Cars vs. Phones: Why Robocars Will Not Save Us

  1. David Levinson
    David Levinson March 13, 2012 at 8:12 am #

    Nice article, some comments below:

    (1) Statistical value of life is about $5.8 million at USDOT, so the cost of unsafety is much higher than you suggest

    (2) The human is either in the loop all the way or out of it. I feel much safer if the human is entirely out of the loop. If we can drive 32,000 fatalities down to 3,200, hooray (you can figure out the economic value of that).

    (3) Yes, liability laws will need to change, we will need no fault insurance, etc. But the insurance companies should be all for this as it will increase safety faster than it reduces premiums.

    (4) Many people cannot or will not ride bikes. Many places are not easily connected by bicycle. While surely there can be more bikes, that cannot solve the general problem of moving people and goods longish distances. If we can automate this mundane task and increase safety and productivity, all the better.

  2. Bill Lindeke
    Bill Lindeke March 13, 2012 at 8:23 am #

    Thanks for the response! I haven't studied this issue a ton, but here are my thoughts…

    1) Yikes.

    2) Robot cars haven't gotten people entirely out of the loop, though. Drivers have to remain somewhat engaged, at least at the beginning and the end. (Though it scares me of at any point, the driver is asleep.) OTOH, transit gets people completely out of the loop. And it's already doing it.

    3) What is "no fault insurance"? Is this possible in a litigious society with private companies making our automobiles and its software? (Think of the case of the Koua Fung Lee, the St Paul driver who had the "sudden acceleration" gas pedal dispute with Toyota.)

    4) Bikes are just ONE possible solution, one that I'm interested in seeing grow. Holland proves that many people CAN ride bikes, as many as half of an urban population. But add in to this picture electric scooters, walking, transit, car sharing, hell even segways, and you begin to get a system of (primarily) nonmotorized transportation. All this is cheaper, slower healthier, and has less VMT than a system of solving congestion that would seem to only encourage people to drive farther (in the long term).

    • Bill Lindeke
      Bill Lindeke March 13, 2012 at 8:28 am #

      I want to make clear that I find the whole technological angle about robot cars amazing and intriguing. I'm just skeptical about the details in the real world, PLUS I'd rather see other things become priorities.

      Why not have trains between our major cities that go 100 miles per hour like they did in 1920? That would be amazing enough for me at this point.

      • Roger Williams March 13, 2012 at 8:51 am #

        I am totally with you there!

        However, I do see robocars as a stopgap, or at least a compromise between "personal freedom" and congestion. Not to mention and wasted time commuting!

  3. Reuben Collins
    Reuben March 13, 2012 at 8:29 am #

    Deploying robocars on access-controlled roadways should be relatively easy – even if only some of the cars have the technology (robocars should be smart enough to detect cars that either don't have the technology or whose technology is malfunctioning and react accordingly). I look forward to this.

    The real challenge will be to deploy robocars on streets that are not access controlled – where we should reasonably expect to see bikes/peds/skateboarders. Robocars should be able to detect the presence of a bike in the street, but robocars are unlikely to be able to react to anything darting, and jaywalking pedestrians etc will be much more difficult.

    Will this result in a push to further separate bikes/peds? or to further control access on roadways?

    On a related side note, I have never understood why we don't use robocar technology to make speeding impossible. Why don't cars automatically limit themselves to the speed limit of whatever road you're on?

    • Roger Williams March 13, 2012 at 9:08 am #

      I would assume that robocars WOULD be able to react to "darting", by applying brakes much faster than a human would react. There would also be more peripheral sensors on the car to detect events far more in advance than a human could.

  4. Andrew March 13, 2012 at 9:51 am #

    The caption on the last photo made the whole article for me. 🙂

  5. Bill Lindeke
    Bill Lindeke March 13, 2012 at 10:00 am #

    This last question is a great one! If the problem is safety, a solution like this would be a no-brainer. That seems a TON simpler to implement than robocars, but would probably have 30% of the effect (or something).

    Of course, it would be the least popular policy since the …

    … actually, I can't think of an actually existing law that would be less popular.

  6. Xan March 13, 2012 at 11:49 am #

    The irony of the photo is that in order to get the picture, you were doing the same thing, playing with your phone/camera/computer. Now, if we can only perfect one-handed sushi…

    Of course robocars don't solve much of a problem because cars themselves are the problem. Taking the driver out of the equation is not the answer. Taking the driver out of the car is.

    Though I must say, I never made the connection between Google™ and the robocar until reading this. Now it all makes sense. Driving distracts people from the internet, duh. I think it would be more worth their while (and money) to invest in a carless future rather than a driverless future. Though that would leave people less time to listen to talk radio and the people there telling them over and over why they should hate people who want to get them out of their cars.

    • Bill Lindeke
      Bill Lindeke March 13, 2012 at 3:36 pm #

      in my defense, I was NOT playing Angry Birds on the bus. I was paying attention to the world around me, which was (in this case) a bunch of people riding the bus while staring at their iPhones. only THEN did I take a picture of it, and only for the purposes of this article.

  7. Brendon Slotterback
    Brendon March 13, 2012 at 3:45 pm #

    Google's car has been operating in human traffic for over 200,000 miles. It has only had two accidents, both of which occurred when a human was operating the vehicle. These cars will not require access control or special infrastructure, they'll work with what we already have, darting kids and all.

    • Bill Lindeke
      Bill Lindeke March 14, 2012 at 6:45 am #

      what about when they don't?

      • Bill Lindeke
        Bill Lindeke March 14, 2012 at 6:48 am #

        i just mean, 200,000 miles is good and all, but the US VMT was 3,000,000,000,000 miles in 2010. there are going to be accidents.

        • Brendon Slotterback
          Brendon March 14, 2012 at 10:15 am #

          Yes, there will be accidents. Probably far fewer than with humans at the wheel. I don't have an answer for the liability question, but my guess is, it would roll up to the manufacturer. They get sued now when things go wrong (unintended acceleration), so if they ultimately "control" the driving through designing the computer, they would be liable in the event of accidents. Of course, they will want those to be minimal.

  8. Julie Kosbab March 14, 2012 at 5:46 am #

    Despite the conflict presented by distracted driving, another challenge of the Robot Car is the mythos of driving. Driving is supposed to be a pleasure, a delight, etc. etc.

    It doesn't typically work that way, but there are a huge number of people who would resist automation in driving.

    I do think there is promise for some types of automation, particularly as it helps people with mobility challenges. Much as hand-controls and the like have helped some disabled folk, additional "robot" features are especially appealing for those with conditions like missing limbs and seizure disorders, and can enhance independent living. That's a niche, of course, but it's an important niche that contributes to lesser costs and greater independence for affected individuals, and thus certainly worth discussing.

  9. Clay March 19, 2012 at 2:10 am #

    Okay, so figuring out the legislative morass will be difficult for robo cars. Sure, I agree, though I think the only real hurdle is getting politicians to let the robo cars on the road in the first place.

    For that issue, I imagine both Republicans and Democrats will get on board with allowing them after they prove themselves in Nevada.

    The liability issue is difficult, but the $5.8m deaths can be factored into the cost of the cars if the manufacturers end up responsible.

    On the other hand, there is the idea of adding pedestrian, bicycle, and mass transit infrastructure.

    Currently, most Republicans appear to be against spending money on any of those things. Not because of how much they cost (outside of mass transit), but just because the idea offends them (e.g., cars pay lots in gas taxes, and bikes pay none at all, so why should bikes get lanes? The fact that cars are getting massive subsidies and bikes almost none hasn't made much of a dent in that argument. And they can always fall back on not wanting to subsidize spandex-wearing, road-blocking Lance wannabes. Or that only insane people get around by bike/on foot anyway. Or that mass-transit riders should pay the full cost of their ride.)

    Even Democrats aren't particularly sold on spending lots of money on mass transit. Or on rebuilding the road structure to make it easy to get around on foot or by bike.

    That, to me, sounds like a tremendously difficult legislative problem. Much more difficult than the robocar problems.

    Finally, there may be two paths before us, but I imagine we'll take both. Cities will get more pedestrian and bicycle facilities, and more of the crash-preventing technologies will make their way into cars, even if the full robo-car doesn't happen.


  1. Why urbanists (and others) should love the coming of the robot car (Part 1) | - April 24, 2012

    […] been written on the robot car, but’s own Bill Lindeke posits that they “will not save us“, citing three problems.  While I agree that we probably can’t be saved (from what […]

  2. Why urbanists and others should love the coming of the robot car (Part 2) | - May 24, 2012

    […] this second and final post in a series, I’m responding to Bill’s reservation about robot cars.  In my first post, I responded to Bill’s objections to robot cars.  Now […]

Note on Comments welcomes opinions from many perspectives. Please refrain from attacking or disparaging others in your comments. sees debate as a learning opportunity. Please share your perspective in a respectful manner. View our full comment policy to learn more.

Thanks for commenting on!