"Self-driving cars are the answer.
But what is the question?"


Dan Hill Opinion self-driving cars

Opinion: after five cyclists die on London roads in nine days, Dan Hill asks whether driverless cars would make our roads safer, or whether dependance on "an algorithmic organisation of society" would only encourage complacency.

Each morning I cycle the wrong way up a one way street, on my way to Fabrica from the old town of Treviso where I live. This is after riding up on the pavement for a bit, and then dodging down a few narrow pedestrian-only lanes. I am not alone in this relentless recidivism. In fact, everyone does it. I even cycled past a police car going the wrong way today, which perhaps provides us all with a kind of collective immunity.

Like many aspects of old Italian towns, this just works. Everything is constantly, steadily, if slowly, moving in all directions. It's like an emergent organism, highly aware of all of its component parts and its overall performance, with hundreds of people working as individuals - and with a vested interest in their own safety - but also highly attentive to their role as part of a wider ecosystem - and so intrinsically aware of the safety and movement of others too.

This attentive mode is essentially the thinking behind the famous "shared space" principle of traffic planning, pioneered over the last few decades by the revolutionary Dutch planner Hans Monderman. It holds that the safest way to design an intersection, for instance, is to actually remove all curbs, traffic lights and signs - and the data backs it up.

I related this idea to city officials in Dubai on Tuesday morning, who looked on in amazement at YouTube videos of shared space systems in the Netherlands. Medium-density intersections, gloomy in the misty Lowlands light, but with a constantly moving parade of bikes, cars, buses, vans and pedestrians. No external controls, fewer collisions. It feels effortlessly better than the frustrating stop-start delivered by most contemporary traffic planning.

Add traffic lights to intersections and you get accidents, as drivers in effect outsource their decision-making to software. In this sense, orthodox traffic engineering makes people less safe and holds cities back from moving to a more civilised condition (take note, Boris Johnson). It's a lesson for urban planners generally, particularly around the unthinking use of technology that we sometimes hear within the "smart cities" movement. Do we want active, engaged citizens taking responsibility for the way their cities work, or passive citizens who outsource their decision making to algorithms?

My favourite Cedric Price line is "Technology is the answer. But what is the question?" We jump on particular technologies - those that gave us car-centric mid-century urban planning, for instance - without stepping back to consider what the real question might be, and hence uncover a wider range of possible solutions.

Listening to the latest pronouncements from Google, Tesla et al about self-driving cars I have similar concerns. While the self-driving car movement says it wants to reduce the horrific number of traffic accidents by removing the element of human error from driving, they may still be a solution in search of a problem.

Here we see such companies are not actually interested in genuine change, for all their bluster about "radical disruption". Self-driving cars are a sticking plaster over existing conditions. They actually reinforce the Californian Ideology that underpins today's mobility problems: suburban sprawl, based around the possibility of lengthy car-based commutes, in turn predicated on a highly individualistic view of society. It is an entirely conservative move. Self-driving cars provide a way of changing the veneer of this system, as no-one is brave enough to suggest changing the system itself. They replace who, or what, is holding the steering wheel, but not the underlying culture that contributes to mass depression, obesity epidemics, climate change and economic crises.

Self-driving cars may be safer - though as Bryant Walker Smith of Stanford Law School's Center for Internet and Society points out, the jury is still out on whether people will actually want them."People are not comfortable with robots killing them," he says. Imagine the legal implications of such an accident. Who is legally responsible? The coder?

If we are interested in safety, the condition I see from my bike every day, or in the Netherlands, is clearly safe. And it uses no new technology; it is a framework for participation amongst its users, and thus it works, as people take responsibility.

Self-driving car advocates might say "Well, this shared space thingy might work in cuddly social democracies like the Netherlands, but not a real car culture like the USA or Australia." But a) recall I live in Italy, and b) remember also Copenhagen's decision, against the tide of orthodox urban planning in the mid-60s, to build around pedestrians and bikes rather than cars. 40 years later they enjoy a highly coveted urban environment. This was no accident, or culture simply playing out, but instead active choice. Every year you delay making such choices, because they are seen as not part of today's culture, you delay genuine change.

The real way to prevent accidents would be to have fewer cars on the road, not just the same number with different control systems. But is the car industry really going to suggest that? Self-driving cars may move traffic a little more efficiently, but the laws of induced demand suggest that the supply of cars might also increase to counter any such benefits.

Few industries could get away with as much blood on their hands as the automobile business does. That we are prepared to expend so many lives - 1.24 million killed each year on the roads, and who knows how many other lives ruined - for the sake of our freedom to drive to work is fairly objectionable.

Even if we don't have accidents, research indicates that lengthy commuting leads quite simply to dying sooner—that is, after becoming fatter, sleeping worse and getting divorced. Happy happy, joy joy.

Equally, it is in the interests of technology industries to propose technology as a solution. We don't spend enough time thinking through the impact of an algorithmic organisation of society. This particular issue was explored in designers Dunne & Raby's excellent United Micro Nations exhibition at London's Design Museum recently, with depictions of an entire society based around algorithmic mobility alongside critical questions of the socio-economic structures it might generate.

And when Tesla, as a hybrid of both sectors, say they are interested too - in Elon Musk's casual, can't be bothered, am I Tony Stark "Well maybe I'll just make one of those, then again maybe I won't (yawn)" kind of way - it doesn't really change the situation of massive resource waste within a car-oriented culture.

Again, this is the inherent conservatism within the Californian Ideology - they actually don't want to change the socio-cultural patterns that they have done well by. To them, technology enables them avoid talking about changing an unsustainable lifestyle. They want to have their cake and eat it. And then get fat.

Yet imagine the possibilities of a city oriented around people living closer to their work and play, and so built around cycling, walking, quality public transport and a massively reduced number of electric cars for individual errands. It doesn't exactly have the airbrushed sheen of Google X, but it would be a city with a lower carbon footprint, healthier people, safer streets, more frequent social interaction, better air quality, quiet enough to hear conversations, to hear birds and to build lighter, more experimental building envelopes, with a higher economic performance through serendipity, agglomeration, richer mixed-use land use, and with increased citizen engagement in the city itself. The benefits are virtually endless, and few are even addressed by self-driving cars, never mind achieved.

Yet as we will always need some cars, they may as well be self-driving. We might then get over the absurd idea of people driving themselves to work or to play. What a waste of time, space, energy and cognition that is! It's a blip in human development that we may look back and laugh about. Or cry.

For all the emotional appeal that cars are associated with, most tasks do not involve an Audi TT gliding around a bend in the Dolomites, sadly, but glum, mildly desperate, slogs through dreary clogged arteries. The FT reports that "the average American spends 38 hours a year stuck in traffic. Cars spend more than 90 percent of their lives idle." And think of the spatial haemorrhage involved in parking space, which is somehow generally empty and unusable. Around 81% of Los Angeles's CBD is parking lots. Imagine the economic potential in actually using that space.

The allure of cars, albeit a fantasy most of the time, is significant though. They are wonderful machines, after all. So why not reframe them as "something for the weekend"? Around freedom, excitement, identity, at your leisure.

And here we get to the real potential of software for mobility - in enabling car-sharing. Given the extraordinary waste of life, carbon and space involved in an expensive product spending most of its life doing nothing, why not really reduce the number of cars required in the first place, via mass, distributed car-sharing schemes?

You choose the vehicle fit for your needs at that point, thus reinforcing the idea that mobility is a bespoke, mass-customised on-demand service shared across bike-sharing, public transport, and through shared self-driving cars for those times when you really need one.

A Fiat Multipla for the trip to Ikea; a Smart car for being picked up from the doctors; a BMW to visit that far-flung fabricator's plant. And drivers can actually book that Audi TT for a weekend jaunt in the mountains. Or maybe a 1969 Ferrari Daytona, why not? The glossy car ad, predicated on the joy of driving, comes within reach by subscribing to a car service, rather than an ownership model. Zipcar++

Software-enabled sharing is far more radical than simply software-enabled driving. We have seen how bike-sharing schemes are beginning to redraw our urban fabric. We can see the growth in the community garden movements. We can see how shared space systems creates a safer, more engaged way of moving around. Self-driving cars have none of these dynamics, simply using software to reinforce what are actually pre-internet ideologies.

Folding self-driving systems into car-sharing schemes, as part of a wider rethink about how we live together in cities, however? I could share that vision. So again, what is the real question that suggests self-driving cars are the solution?

Image courtesy Google.

Dan Hill is CEO of Fabrica, a communications research centre and design studio based in Treviso, Italy. He is an adjunct professor in the Design, Architecture and Building faculty at University of Technology, Sydney, and his blog City of Sound covers the intersection between cities, design, culture and technology.

  • Speed

    “Add traffic lights to intersections and you get accidents, as drivers in effect outsource their decision-making to software.”

    The accidents happen when drivers ignore the decisions made by the software.

    “Cars spend more than 90 percent of their lives idle.” My toilet spends more than 90 percent of its life idle. Worse, I have three of them in my house. Perhaps I should get rid of two of them and install some scheduling/sharing software to more efficiently use the one remaining. And while we’re at it, why not have all the apartments on one floor share a kitchen instead of each having its own that sits idle much of the time?

    Sharing is fine and should be encouraged when it makes sense. But it is up to each individual to decide. Do I want to own that book or share the one in the library?

    • George Joseph Lane

      But in this case it does make sense. Your toilets are very cheap and spatially efficient, your car isn’t. Likewise with books, those that you read often it makes sense to own, those that you may need as a reference for a business report once in your life should be shared.

  • BriH

    Some sense at last regarding urban planning & traffic; I will forward this to my city planners, but I am not holding my breath that things will change anytime soon. Too many vested interests methinks!

  • sean

    Why don’t we just get driverless bikes? I think that cyclists tend to have much less regards for the law and flow of traffic as a whole. I understand that in most areas the “driver” is responsible for the “cyclist” but feel that it is here therein that the problem exist. Also, accidents will always happen.

    • zd.enko

      Straight to the point!

      Article which begins with a description of one cyclist defying the law. I wonder what would happen if car drivers would decide to ignore one way signs? Or traffic lights maybe?

      Well known shared-space systems have only been tested in small towns of some European countries, but I can’t help but wonder would their statistics be so positive in big cities where some person, unfamiliar with the neighbourhood, decides to rush through the intersection without stopping in the middle of the night.

      I believe, as a pedestrian and a car driver, that bike drivers are the biggest threat in traffic with, in most cases, totall disrespect for the former, and not following the traffic rules when interacting with latter.

      Returning to the subject of the article, self-driving cars may possess threat to the bike drivers, but only because bike drivers are the most unpredictable element in traffic and biggest threat to themselves.

  • Paul Godsmark

    I put forward the argument to the TRB Workshop on Road Vehicle Automation that many of the Urban Planner’s goals actually find great synergy with the beneficial attributes of the fully automated car. On my (poor quality!) poster presentation I showed that the urban planner desires safer, more liveable streets that are conducive to cyclists and pedestrians, that will attract businesses and commerce and have less pollution – and more space! Which is exactly what can be achieved if we close off an automated vehicle zone (AVZ) to human drivers and use ultra-lightweight electric shared-fleet automated vehicles for moving people (and even goods) around.

    Because they are shared vehicles there is a considerable reduction in parking requirements as they are constantly being allocated new rides – so curbsides can be recovered for communal/commerce use. Parking lots and structures can be reclaimed – but municipalities and developers will have to fight over their use! (raise money, or benefit the community!) – good policies are needed.

    Because the automated vehicles (AVs) are so safe it is much more difficult for the peds and cyclists to have crashes with them – and the priority level given to peds/cyclists can be programmed into the AVs. On street festival days then peds/cyclists have absolute right of way for instance.

    The maximum potential benefits of AVs are unleashed when we have close to 100% AVs – and they improve just about every aspect measured by the Mercer Liveability Index – which will make the Mayor and Councillors very happy if their city suddenly moves up that Index – as this is very good for business, investment, social aspirations etc.

    My poster can be found here: http://autonomous-vehicle-impacts.blogspot.ca/2013/07/why-automated-vehicle-zones-avzs-might.html

  • Dan

    It’s an interesting way of putting the technology problem/solution but then again what problems were the TV or iPhone or even car solutions to? Now we can’t live without them.

    The driverless car will be to cars what the iPhone was to telephones I believe. And nobody is saying that we will get rid of bikes or fix every other problem in the world but this will part of making life easier. Or maybe we will all end up fat and being driven around like the people in Wall-e, maybe or maybe not. Who knows?

  • Steve X

    Self-driving ‘vehicles’ already answer a question. Namely, how can we have fewer people on mines?

    Rio Tinto are already doing this.

    They already answer the question of how can we have fewer people in dangerous, expensive environments.

    The US military will have self-driving supply vehicles before the day is out.

    Then they answer the question of how we can we have cheaper buses? A bus driver costs about as much in total costs every year as a self-driving system does to buy.

    Then they answer the question of how can I have something with the utility of my car cheaper? My car on a day I drive to work is used for about 50 minutes. For the rest of the day it is wasted. If the car drove itself it could take 3-4 people to and from work and have the same utility value to me.

    So, I hope I own my last car. Because the final question I want answered is how can I my garage back and cease using this expensive piece of real estate as a storage dump.

  • DrPlokta

    Since you seem to be short of them, here are a few questions for you:

    How can we make personal transportation safer?
    How can we let people in areas of low population density work while they’re commuting?
    How can we let old and disabled people keep their mobility?
    How can people get home after an evening drinking in a country pub?
    How can we reduce the cost of buses and taxis?

    • I scrolled to the comments section to write exactly what Dr P has written. The author seems to think that everything revolves around his bike riding. Sorry, but bike riders are a small concern, when it comes to self-driving cars.

      Self-driving cars will be safer, more efficient, and cheaper, than human-driven cars, as has already been demonstrated by Google’s prototypes. This is true for all of us on the road, walking, in cars, and on bikes. It is so self-evident I am surprised the author ventured the question.

      • David Sucher

        Uh…did you read Dr Plotka’s questions? They don’t seem to me to have anything to do with bicycles whether for or against. In fact the answers to his questions are all pro driverless car.

  • Dan

    This article was on my mind for a while and here are some of my thoughts.

    Driverless cars would probably compete with buses, trolleys, taxis and other medium and short distance vehicles maybe even trains. A questions to ask would be how would these transportations system be affected? Cars would essentially become tiny personal buses.

    The next question to ask is how and who will pay to use these cars. Will they work like taxis and pay for every ride or will they become a public resource like water and we all pay a tax to be able to use whenever and however? In every case, the biggest effects would be to make transportation available to lower income people therefore making it possible for people who can’t own a car to get to work anywhere and not worry about public transportation schedules?

    Sure the streets will be safer, old people and everyone else will be able to go anywhere but will that mean. I see cars as becoming specialized train cars that do not need and train engine to pull them. You can get a car for the specialized task you need such as a truck to haul materials or smart for short distance.

    The questions we should be asking is what will we do with the free time we will have not driving? Also what will cars become when they don’t need a driver? Will they become a lounge on wheels? Another kind of bus? A hotel on wheels? A bar on wheels? What is the future of the car and its effects on society?

    The solution (driverless cars) brings about more questions for me.

  • Calgaryairportlimousine

    Interesting concept but you will be asking people to give up control and many will not be willing.

    • David Sucher

      Then they won’t use the driverless car. Unless of course it turns out that accidents are primarily caused by high-speed or drunken or otherwise unskilled HUMAN drivers and, starting with the Interstate Highway managers, no car will be allowed on the freeway unless it’s in driverless mode.

    • RaymondPink

      It’s not about accidents, the technical efficiencies, and energy savings. This is about people losing rights to choose.

      This is about Government controlling our lives. I believe there will be no choice once the majority of vehicles on the roads are driverless. That day will be when the minority must get off the roads or comply because of compatibility in the name of safety.

      That is also the day robots take over your life. Thats the day I say f*ck all this technology and head for the hills.

  • Michael S.

    I live close to the Netherlands (Belgium) though I’m American and have driven there many times. What they have are roundabouts, where everyone knows that entering traffic yields to traffic already inside the roundabout. Also, everyone knows that pedestrians / bikes using a crosswalk have the right of way. These rules are ingrained. They also have plenty of traffic lights and signage on a lot of roads / streets, so I’m not sure what the writer means. I love the car-sharing concept though.

  • Jack

    The premise of this article is motivated more by dedication to socialism and centralised planning rather than urban planning considerations. Driverless vehicles will be another nail in the coffin for socialist living and crowded filthy inconvenient public transport.

    Living in a low-density environment and owning multiple cars is not “selfish”. Some people just don’t want to live around others. Some people don’t feel this urgent need to “engage with community”.

  • David Sucher

    What is the point of advocating “for” or “against” the driverless car?
    1. The technology will either work or it won’t. If it doesn’t work, we won’t use it. If it does, then we will solve some problems and find a host of new ones. That is life.

    2. Why should not the driverless car be able to negotiate shared-spaces?

    It would certainly be worthwhile that driverless cars meet certain standards to include being able to use shared spaces. Frankly I don’t see much of a problem from what I have read but let the facts show us one way or another.

    3. I think that the author simply doesn’t like cars that dominate human settlements. I share that feeling to some degree. But I think that the driverless car has the potential to do a great deal of good for better and more livable cities. Just imagine a system which combines the driverless car and Cars-2-Go (or similar shared auto systems.) We’ll see a reduction in required parking, more space for infill etc at the same time as we see potential for greater suburban expansion. On balance I see more good than bad, with the exception of issues surrounding surveillance/NSA.

  • Tedd

    In effect, this article asks the burning question, “Should you walk to work, or bring a lunch?” There is no sense in which autonomous vehicles are contrary to, or even compete with, the ideas promoted here. The author is merely attempting to piggyback unrelated ideas on a currently popular and controversial subject.