Could Google’s Self-Driving Car Kill Public Transit?

Post by Steven Dale

Google's Self-Driving Toyota Prius. Image via Endgadget.

Last week at the TED 2011 conference, Google unveiled their self-driving car. And while an interview and test drive with CBS news was both fun and informative to watch, the issue of the technology’s impact on public transit was never addressed.

If the self-driving car becomes a reality in the next, say, 20 years, one has to ask the very valid question of what purpose, truly, will public transit serve in the near future?

Will it kill public transit as we know it?

That may seem alarmist, but it’s certainly worth asking and answer to the above question is anywhere along the spectrum from possibly to probably. But whatever the impact, it is a game-changer the likes of which public transit has never seen.

Last fall I laid out what I thought was a fairly convincing scenario for how the technology might be adopted within our lifetime and the requisite comments and disagreements ensued. So rather than just assume an inevitable that we can’t possibly predict, here are 10 reasons I think public transit could have a very rough ride if this technology gets off the ground:


ONE. It’s an easy technology to spread virally.

Technology that requires the participation of others to increase its usefulness propagates rapidly. Fax machines, email bank transfers and facebook are only useful to the user if other users exist as well. Google’s cars are to be “learning computers” that improve with time. The collected wisdom of millions of these cars will be infinitely more useful than a couple dozen. People who initially purchase these vehicles are likely to pressure their colleagues and friends to purchase them as well.


TWO. You can work while you drive.

One of the single greatest advantages transit has over the private automobile is that it allows riders to be productive while another individual does the driving. Granted, transit agencies rarely capitalize on that advantage, but let’s assume most agencies did. That competitive advantage vanishes when your driver is replaced by a computer.


THREE. Safety.

Public transit kills and injures thousands of people per year across North America. Not necessarily riders, but pedestrians, cyclists and other motorists. And while the initial argument in favour of driverless cars will be the elimination of accidents caused by other private vehicles, it won’t be long before people start to notice that the driverless cars are actually safer than public transit.


FOUR. Lobbying Power.

Google is richer than God and Donald Trump combined. One thing that could prevent this technology from developing is not having active co-operation from a municipality friendly to the idea. But you can bet that with the power and money Google wields, that’s unlikely to happen. Google is bound to find a willing municipal partner willing to a gamble on technological innovation.


FIVE. The Bay Area.

See number 4, then apply it to San Francisco and Silicon Valley, Google’s home territory. The Bay Area is technological innovation. They will want to be first.


SIX. It’s A Consumer Item.

A driverless car would be consumer-oriented and customizable to the needs of every individual and family. Public Transit has yet to shed its image as a charitable item and position itself as a pleasant, consumable good. Whether right or wrong, consumable goods are simply sexier in the eyes of the public than social amenities. What happens when Apple launches its inevitable cute, little, white iCar?


SEVEN. Public Transit is likely to adopt the technology, too.

Let’s face it: Bus drivers are expensive. So is scheduling. So are rail tracks and subway tunnels. If a transit agency recognizes that they can offload the cost of maintaining all their track infrastructure onto the local roads department, you can bet they’ll do it.


EIGHT. Non-drivers will become car-riding non-drivers.

There’s a large segment of the population that doesn’t drive. Many of those people don’t drive by choice, but others are prevented from driving for any number of legal, social, economic, or health-related reasons. But nevertheless, the stigma of not driving is likely to remain. This technology eliminates that problem. At first only licensed drivers are likely to be allowed to “drive” these vehicles, but if the technology proves itself, those restrictions will be lobbied away in order to open a massive new market for the private car.


NINE. The car as office and living room.

Even more than the current crop of soccer-mom SUVs that have more amenities than a Manhattan condo, a driverless car will become an extension of one’s entire life. Travelling is a disruptive process because it forcefully disrupts our lives and routines and puts humans in contact/conflict with other unknown strangers. A driverless car would eliminate that disruptive process. Anything you can do while sitting suddenly becomes something you can do in Google’s vehicle – in an environment customized to resemble that which you already know. The productivity and efficiency such a development could create would be one of the strongest arguments for people to adopt the technology.


TEN. People will want to ride these.

Admit it: You know you want to already.

Want more? Purchase Cable Car Confidential: The Essential Guide to Cable Cars, Urban Gondolas & Cable Propelled Transit and start learning about the world's fastest growing transportation technologies.

Want more? Purchase Cable Car Confidential: The Essential Guide to Cable Cars, Urban Gondolas & Cable Propelled Transit and start learning about the world's fastest growing transportation technologies.


  1. Even self driving cars can still get stuck in traffic, and they still need parking, and they don't make neighborhoods more walkable and attractive... I'd rather ask whether the self driving car will help kill individualized transportation. Automation like this makes demand-responsive transit possible; and trains will always be used as trunk lines, because they are much much more energy efficient, require much less maintenance (and have lower operational costs per passenger if done right).
  2. it's still more expensive then bus fare... wonder if they will be easier or harder to steal. computer hackers are thieves of the future.
  3. @ant6n They don't need parking if they're used as a public system and can automatically go pickup someone else each time someone gets out. @frankie g The big operating cost for transit is labour. A whole fleet of driverless cars needs perhaps only a handful of dispatchers and might actually be cheaper then buses. As for your tenth point, Steven, I don't want to. I expect many people will, but I already hated priuses before they became robots. I'll stick with my manual shift and non-ABS brakes and un-assisted steering thank you google.
  4. @3: Compared to individual cars, collective transportation has basically no deaths.
  5. Telecommuting is going to take over so there will be less cars on the road. When people realize that they don't need their car as much, they will try to sell them but everyone will be selling. We will have vehicles rusting in people's yards. The government will then subsidize transit and give incentives to people to scrap their cars. The transit companies will of course go autonomous with its buses. They will also own a fleet of the autonomous cars to get people to and from the transit lines.
  6. Point 7 is very right. But that technology won't kill public transit. I think public transit will in near future even be more important than it used to be ever. With the gas price and the price for a good car, the missing technology of getting around for a thousand km by one battery charge - a lot speaks against the killing of public transit. @Erik: "They don’t need parking if they’re used as a public system and can automatically go pickup someone else each time someone gets out. " I wouldn't be so sure about that because people still want to own their own car. And who wants to share a Mercedes or any kind of exciting car? I'd share a 20 year old wrack, but that's it. So, what could be done is to have a public transit fleet like taxis, maybe a big bigger but smaller as common buses and they will handle the jobs they always used to - because like ant6n said "trains will always be used as trunk lines, because they are much much more energy efficient, require much less maintenance (and have lower operational costs per passenger if done right)." Yes I want that technology (let's say to avoid accidents in seconds-sleep, or for older people to be able to "drive" a few years further, for me to maybe even really text with my mobile phone while driving - or simply to drive without driving) I'm quite sure cars will always be able to be steered old fashioned.
  7. I see there a big mistake to think. You have car jams now because too many people want to travel with their car at the same time to the same places. And there is no difference if you travel alone with your own car or alone with an automated people mover. Perhaps you save a second car in the family, but you don't save the car jam in the morning or evening. But you can transport more people at the same space with an urban transportation system. If you want to transport 4,000 passengers/hour, you need: 3,077 cars (1.3 persons/car) or 1,000 cars (4 persons/car) or 42 buses (50 persons on seats and 45 standing) with 42 drivers or 167 gondolas (24 passengers) with 0-2 persons/station, every 22 seconds These 3,077 cars need space, = 17 km traffic jam, more at high speed (length of the car + distance) x 3,077 If you transport one person you need the weight of the vehicle of 850-1460 kg at a car (1.3 persons/car) (and you need energy to transport this curb weight) or 150 kg at a gondola (24 persons inside) Do you really think, (automatic) car traffic has a future ?
  8. @ Gunther, "You have car jams now because too many people want to travel with their car at the same time to the same places." I think the one thing you miss here is that with a swarm of computer controlled vehicles you would have vehicles running closer together, without the need for stoplights and a more efficient use of total road space - rather than the current scenario of most roads being completely empty and every arterial road and highway packed to overflowing. And yes, I'm convinced "automatic"/self-driving/driverless cars have a future. It appeals so deeply to basic human psychology, it can't possibly fail. Will that mean the end of public transit? Who knows.
  9. "Do you really think, (automatic) car traffic has a future ?" I also see a bright future. And most traffic jams are caused by human nature, not by the system itself. Everbody does stop and go differently. Some people react to late when a traffic light turns green, others stop the engine while waiting for a green light, others do play with mobile phone or radio. It's a phenomenon why traffic jams exist on highways (lots of them without accident). And I'm also convinced public transit will remain. Cities like London won't survive without it.
  10. These 3,077 cars need space, = 17 km traffic jam, more at high speed (length of the car + distance) x 3,077 Distance is 3 m, sum of distances = 9 km Your car is 4 m long, du = 12 km. You need space of 21 km to transport 4,000 passengers. The system of Automated Peopla Mover Cars functions with 2 cars on a test track but not with a full track with cars in 3 m distance. What happens, if one car wants to enqueuing into such a queue of cars with 3 m (or 3 cm ) distance? One car is stopping and all cars behind... Or I have to wait and to wait and to wait and the cars behind of my car. Car jams are the result of too many cars on a road, now and automated drived 2050. When I come with my car to a destination stop and it is full (too many cars at this part of the road). Then I have to make a circle like an airplane again and again or I have to wait and make a jam? When automated cars with distance 3 m come from 4 roads to a crossing = a roundabout, the cars automaticall mixes? The distances decrease to 3/4 m ? And if all want to drive to one direction... There is no difference if you travel alone with your own car (and make a jam) or travel alone with an automated people mover (and make a jam).
  11. Guenther, have a little faith. Imagine how computers looked like 50 years ago, what they were capable of 25 years ago and what a common mobile phone is able to do at this time. Mobile phone, camera, music player, navigation system, media player - all in one. Cars turned out pretty much the same and lots of other things as well. In case of the theoretical part within the ongoing topic: The latencies within that computer-assisted approaches zero: means all cars will be able to react like a chain. when the first one brakes, all the others do the same at the same time. no stop and go. and let's say that technology is just used to avoid accidents. have you ever been in an accident or know someone who did? almost all of them are caused by human failure! I don't want to be too specific, but 3077 cars are not automatically a traffic jam. So, with an average of 1500 to 2500 cars per hour and track a common double tracked road would do the magic to make enough space for the traffic (and nevermind the length). More guarantee upon a more fluent traffic though would give the "computer-assisted driving thing". Here's an article about it: http://de.wikipedia.org/wiki/Stau_aus_dem_Nichts (The article is about the topic of traffic jams out of nowhere and the causes) And that is what google is saying right now: it's a technology to reduce accidents, make traffic safer (and maybe in future cars will be driven driverless - but they do not promise that, they just say that might happen, maybe)
  12. It is a problem of not enough space. If the number of cars increase at one kilometer space (because of access roads), logically the distances between the cars decrease. If your safety distance to the car in front of you decreases, you reduce your car speed. And so the traffic flow becomes slower and slower and slower. And at the next access road or at a stupid braking action the traffic flow is at 0 m/s. I can observe it, when I am stupid enough to drive my car at rush hour in the morning to the next bigger city.
  13. I want to see it in reality, if you have a queue of cars with 4 m distance how a car from access road is planted/fitted into this queue. The big-brother-computer knows the arrival of this car, opens a gap with 10 m free space and the car whooshes inside? This I want to see.
  14. Well in this case we agree. We'd like to see it too. =)
  15. using the argument of a 4m car is pointless because we all have seen that as technology progresses it also shrinks. therefore as cars become smarter they too will become smaller and lighter. that solves one problem. next, as sean pointed out, telecommuting is on the rise and with it the need for people to move about on such a daily basis is diminishing. if you use your car to get to work, and you no longer need to go somewhere to actually work, you won't drive your car, at least not at that time for that trip. that more or less solves the issue of rush hour. so, let us consider the "smart prius" a mock up. as the brain of the car develops, the body will adapt and follow suit. we can either wait and see this happen before our eyes, or help it happen along with input and insight.
  16. @erik im not referring to the cost of a bus vs the cost of a car, but rather how much its going to cost me, the lonely little dude who buys transit tokens not entire fleets of buses. i could care less about the cost of the full scale operation, but at the moment, since the cost of a car is out of the question when i need to get across town there better be a bus to get me there!
  17. finally – the words "auto" and "automobile" will have real meaning
  18. Thanks for the suggestions you have contributed here. Something important I would like to express is that computer system memory requirements generally increase along with other breakthroughs in the know-how. For instance, when new generations of processor chips are made in the market, there is usually a similar increase in the shape preferences of both computer system memory in addition to hard drive space. This is because software program operated by these processors will inevitably surge in power to take advantage of the new technological know-how.
  19. @off the cuff the auto in automobile means self, not automatic
  20. Not going to telecomute your plumber or auto mechanic job. Parking will be unnecessary, the car can just go back home and park in your own driveway. Driverless cars will be able to tailgate safely, and "draft" the car in front for dramatically increased energy efficiency. Driverless cars will leave the house and begin the trip exactly when you want to, deliver you to the door of your destination, and return exactly when you want to, operating on no one else's schedule. Nobody will ride mass transit because of these things, so we will be looking at a world where there is just 1 transportation task outside of air travel, and that will be to build roads. Fortunately, roads are waaaaaay cheaper and easier to build than, say, subways, so transport will get lots cheaper because of the self-driving cars.
  21. @Dave Have considered that a driverless car that delivers you and then go back home is an enormous waste of time , energy and roadspace ? A system like this could simply multiply x10 the actual traffic and also could make you lose more time than with a simple car . Mass transit exist just because ... you have masses to move from A to B. And , please , have a look to real road construction costs ...

Leave a comment

You can add images to your comment by clicking here.