The Myth of The Self-Driving Car Cure

There’s this mythos spreading about the potential of the self-driving car–that it will bring an immediate end to traffic jams and collisions, alleviate the stress of long commutes from the suburbs to the cities, obviate the need for public transit and for parking, spur urban revival and simultaneously stretch suburban sprawl further, and cure many problems of owning and operating a car–but that’s just what it is: A myth.

94100d05737425018470d6bc9f31299eThe dream of automated, more private transportation is not a new one; Americans have been working on Personal Rapid Transit, systems of vehicles that can be driven independently or can connect to a rail system for mass transit, since the 1950’s with little progress. The technical obstacles, logistics, and (most importantly) costs were just too great.

Self-driving cars have the advantage of more flexible and more capable digital infrastructure over the electric rail PRT relied on, but that systematic independence also increases the overall complexity of a system with self-driving cars. And those complexities are being overlooked by urban planners and suburban state officials alike who herald in the end of traffic, the obsolescence of parking, and the death of public transit.

There are a number of issues with the Self-driving Car Cure that have yet to be resolved, and likely will remain unresolved for years to come.

Liability: When there’s no one in the driver’s seat, who takes the wheel?

Legally, self-driving cars are still way down the road from being something anyone can go out and buy. Only four states (and D.C.) allow self-driving cars, and those rules don’t even begin to cover the bases of liability. Though self-driving cars are designed to avoid collisions, they are not free from error. One of Google’s own autonomous vehicles caused the company’s first collision just last month.

In the future of self-driving cars, who is responsible in the case of a collision? The owner of the car? The car manufacturer? What if the car is owned or operated by a ride-hailing service and it is involved in an accident- does the responsibility for the collision fall on the service, the manufacturer, or the passenger?

All of these questions and more have yet to be answered, but the US Department of Transportation is making headway on that matter. It’s highly likely that liability will be decided based on what class of self-driving car is in question. Buckle up, it gets kind of wonky here.

There are multiple levels of autonomy for self-driving vehicles: the National Highway Traffic Safety Administration first discerned the distinctions in the new technologies in 2013 with its “Preliminary Statement of Policy Concerning Automated Vehicles” under former Secretary Ray LaHood. There are five levels of automation:

Level 0 cars put the driver solely in control of all major vehicle functions like steering and braking, but do include technologies that assist the driver like blind spot monitoring, and automated controls like windshield wipers;

Level 1 vehicles have function-specific automation, in which the driver is still solely in control of driving functions, but the vehicle has technologies that help the driver operate the vehicle, such as cruise control, lane-keeping, and collision warning;

Level 2 vehicles allow drivers to cede operation, “disengaged from fully operating the vehicle,” in limited circumstances to the vehicle systems;

Level 3 cars are capable of fully controlling “all safety-critical functions under certain traffic or environmental conditions.” The line between Levels 2 and 3 are a little blurry, and this is where the self-driving cars from Tesla and major automakers land. These will most likely represent the vast majority of the self-driving market as a result.

Level 4 autonomous automobiles do not let the driver control the operation of the vehicle at all, only letting riders decide the destination and environmental factors inside the vehicle. Think those cute little cars Google is developing. Ride-hailing services like Uber and Lyft are extremely invested in developing these technologies as well. These are what most people imagine when they hear “self-driving” or “autonomous” car, fully independent of drivers. These are the self-driving cars you’re looking for.

In February, the NHTSA gave Google an unexpected clarification on the liability concerning the Level 4 vehicles the company is developing: “If no human occupant of the vehicle can actually drive the vehicle, it is more reasonable to identify the ‘driver’ as whatever (as opposed to whoever) is doing the driving.” That would suggest that the liability in the case of collision would fall on the producer, in this case Google, of the self-driving car’s operating system, which is a huge step. This followed the announcement in January that the US DOT will have a framework for autonomous vehicle regulations ready later this year, providing a foundation for states to build on (or not). Some members of the Senate are also pushing for legislation to require the NHTSA and the Federal Trade Commission to regulate self-driving cars to protect drivers and riders from cyber-attacks, but car companies are pretty opposed, setting the stage for protracted debate.

Google’s Level 4 prototype autonomous vehicle in action.

Most experts in the industry expect that self-driving cars will be road-ready in five to ten years, at the earliest; I’m of the opinion, based on the recent developments of Tesla, Google, Ford, and other companies that this timeline will be much shorter. After all, Tesla already rolled out its beta-feature autonomous driving update to some customers’ Model S’s last year, only to remove it with an over-the-air software update after sufficient backlash to videos posted by Model S drivers of the autonomous driving in action.

At this year’s CES (Consumer Electronics Show), self-driving tech dominated the convention, as they did the year before, and the year before that. Self-driving cars have been in the works for longer than most people realize, but that doesn’t mean they’ll be market ready soon–which is actually a good thing. These policies will give cities, states, and the Federal government a little more time to figure out how exactly to regulate and plan for this technological change. California, Nevada, Michigan, and Florida already allow self-driving cars on the road. In five years’ time, you may in fact be able to get behind the wheel of a self-driving car anywhere in the nation without worrying about the legal ramifications.

Logistics: Will self-driving cars really revolutionize parking in cities?

Those that do buy self-driving cars will still have to store them in parking garages–where else would they go?

Well, self-driving car boosters propose that they wouldn’t need to park, instead they could ferry other passengers about through ride-hailing serves like Uber and Lyft. But that would require folks to own level 3 or 4 autonomous vehicles and to actually want to share their cars–considering Business Insider reported in 2012 that “the massive array of sensors Google has to install in its cars alone costs $250,000 or more,”–that isn’t likely until prices come down. Additionally, temporarily leasing autonomous cars to other riders is dependent on many years of legal groundwork and technological advancement before it would be feasible.
Mercedes-Benz’ latest concept car, the F015, can even drive you and three other passengers to an isolated spot in the desert to dispose of you, probably.

If not, many self-driving cars, for all their merits, will likely just end up in parking decks with our dumb cars of today, which are parked a whopping 95% of the time. There is another options, proponents say: Self-driving cars could also be able to return to the homes of their owners, like newer Roombas do when they’re done vacuuming up our messes.

That sounds fine and dandy at first, but when you think about it, as time advances and autonomous vehicles assume more of the overall automobile market, that would mean that more and more cars in the morning rush to the region’s economic centers through major corridors would be making that trip twice. The result would be increased traffic periods, increased greenhouse gas emissions, and a lot of wasted energy.

Three things in life are certain: Death, Taxes, and Traffic.

If self-driving cars are not shared by their owners and also have less space to park in cities, how does the argument that they will reduce traffic and parking make any sense? Will they just circle the block, never stopping until beckoned by their owners at closing time? Maybe we’ll go back to building those parking skyscrapers New York city had in the 50’s or maybe more crazy designs like these.

On top of that, recent studies of congestion show that just one car in a heavily trafficked roadway can bring traffic to a halt. As explained in the video below, if human drivers behaved as ants do, then traffic would flow much more smoothly. So perhaps self-driving cars will be able to emulate that behavior, but even with all the sensors and computing power crammed into the cars of the future, we’ll still wind up bumper to bumper until there are more self-driving cars on the road than not.

The only way I see this fantasy of cars returning to their owners’ homes without the negative externalities of increased vehicular traffic is if all future autonomous vehicles are electric, level 3 or 4 autonomy, and stay in the far right lane during commutes. By being electric, they will have reduced vehicular emissions, though unless alternative sources of energy expand more rapidly across the nation, there will still be enormous increases in GHG emissions–just not in Metro regions. By being level 3 or 4 autonomous, the vehicles would be able to drive themselves with a certain level of risk aversion to others on and around roads. By staying in the far lane, they will be out of the way of other traffic on highways and travel at slower, safer speeds; I also have no doubt that autonomous vehicles would be able to let traffic on and off better than their human drivers can (at least in Atlanta). That’s a huge set of big “if’s” to make that fiction reality. We’ll see if carmakers can do it.

Land-use: Let’s be realistic–Will people buy into the idea of sharing cars in the ‘burbs? Will they make driving further distances from far-flung suburbs less harmful?

In a car culture like America’s, much less Atlanta’s, people are not just going to give up their cars in droves. In fact, more Millennials today are buying cars compared to a just few years ago during the Great Recession,, even though most of those are used cars. Bloomberg View reports that as a percentage of total consumer expenditure, Millennials still spent half their expenditure as the same age group in 1989 did, and average spending on new cars dropped by over two thirds from 1989 to 2014. The old wisdom that most people buy cars as they age and have children seems to hold true, too, but surveys of younger Millennials and Boomers show that a growing number of them don’t want to own cars at all. The majority still do buy cars because they live in suburban areas where not having a car is an inhibitor to employment and socializing since everything is so far apart. We’ll see if that trend holds up since many Millennials that move to the suburbs don’t really want to live there. Given that data and the fact that the majority of self-driving cars under production are pretty high end models, it doesn’t look like many Millennials will be able to afford autonomous automobiles anytime soon anyway.

Proximity to jobs increased in Atlanta’s suburbs between 2000-2012 than in the inner city. Source: Brookings.

Similarly, demand for walkable places is soaring and changing the face of American suburbs, especially Atlanta–and that demand won’t be met anytime soon. Even as suburbs adapt to become more ambulatory, by and large they are still isolated from transit connections. However, without transit connections, the walkability of these places is mostly a façade, but a bullet point on a billboard; To quote Darin Givens, they are islands of “drive-to urbanism,” in America’s vast seas of suburbia. In places like these, self-driving cars could theoretically supplant transit.

Millennials value options and in many cases consider purchases that used to be essential as burdensome. That’s one reason why transit ridership continues to climb, with 2014 having the highest rates of ridership in the US since 1956. Even though most still need personal vehicles to get to their jobs, it doesn’t look like they’ll want to in the future as they (hopefully) become more affluent. What’s more, Uber and Lyft riders are also more likely to use public transit as well, own fewer cars, and spend less on transportation, a new study of 4,500 ride-hailing service users reports.

But if people aren’t going to give up owning their own cars, will they be willing to share them? This isn’t your parents’ HBO GO password, after all, it’s a car. During my research, I haven’t yet found a study that explicitly asked if Americans would willingly let other people use their personal cars without supervision, much less as on-demand taxis with services like Uber and Lyft.

Self-driving cars, the theory goes, will be able to ferry passengers without supervision, earning some cash for their owners on the side and cutting costs for ride-hailing services by not having to pay drivers a large cut.

But would people in the suburbs be interested in that? “No” seems to be the obvious answer–people that live in the suburbs are typified by the desire for higher levels of privacy, personal space, comfort, and control. Even so, having someone drive your personal vehicle, while you’re not in it and they aren’t really driving, is an unsettling idea regardless of where you live, in my opinion.

On the other hand, if services like Uber or Lyft owned the vehicles or partnered with an automaker, like how Lyft has partnered with GM, it may be much easier to convince consumers from buying cars of their own. James McQuivey, an analyst with research outfit Forrester, told WIRED that “Selling individuals a self-driving vehicle isn’t as game changing as getting people to use self-driving cars they don’t own. The former only adds a new technology while the latter changes the economics of one of the world’s biggest industries.” Considering Google is partnering with Ford and Uber is also working to develop its own self-driving cars, this seems like a possible scenario, in several years.

A few of Google’s autonomous Lexus SUV’s, of which one recently  a collision with a Santa Clara Valley Transportation Authority bus.

The Road Ahead

There’s no doubt about it: the future of self-driving cars is wildly exciting and brimming with potential to make people’s lives better. These cars will be able to circumvent the realities of our current road systems, one day, and they may enable personal and economic mobility not seen since the original boom of the personal vehicle after World War II.

But we need to be realistic and not get caught up on this hype train.

Randall O’Toole, Fellow at the Cato Institute and scholar of the Georgia Public Policy Foundation, argues in two recent op-eds that self-driving cars will bring about this revolution in transportation, replacing public transit and allowing suburbs to sprawl indefinitely.

“Self-driving cars will change the way we think about distance,” he wrote in his op-ed in the Washington Post a few weeks ago. “If people can work, watch movies or play video games with their children on their morning commutes, they will be willing to live much further from work than they do today. This means urban areas will become even more spread out than they are now.”

Never mind the housing market shifting towards denser, walkable communities, the negative health and environmental impacts of long commutes; the absurd costs of suburban growth; and the fact that transit ridership continues to climb–people still don’t enjoy their long commutes. In Atlanta, it’s the #1 issue on voters’ minds, with the average Atlanta region commute hitting 59 hours every year. Atlantans can take comfort in not being in the Top 10 U.S. cities for worst traffic congestion though–we’re only 13th. I’m not sure that being able to watch Netflix on your commute five or 10 years down the road will suffice as a response now.

The mistake that so many pundits are making is in pitting self-driving cars against transit expansion as a solution for future transportation improvements and investments. Here’s looking at you, O’Toole: He also advocates for completely eliminating transit spending for all but “low cost, flexible bus service,” which is literally the same consultation given to American cities in the late 1940’s and 50’s when faced with the decision of supporting their existing streetcar networks or restructuring their cities for highways. Well, we know how that panned out.

O’Toole’s argument is based on the false foundation that transit is as a technology fundamentally flawed, ineffective, and has reached its technical zenith. O’Toole also ignores the economic development impact of rail transit when paired with zoning that permits denser, walkable growth, as well as the popular pleas for each. In focusing solely on (and exaggerating) the potentials of autonomous vehicles, O’Toole misses the bigger picture; he ignores those that cannot afford to wait for self-driving cars and possibly won’t be able to afford them for all their transportation needs. He dismisses the desires of those that don’t want to sit in their cars for hours every day, that want their workers to be healthier and happier, and people that want to save some cash.

All this isn’t to say that autonomous vehicles won’t have a huge impact on American culture and society, they will. That doesn’t mean transit will be left by the wayside. The technologies being developed by the private sector are applicable to the public domain, one reason why the Federal Department of Transportation is investing so much in smart cities and R&D for autonomous vehicles and transit. Just last month, the Urban Land Institute hosted a panel discussion of transit and transportation experts in Atlanta, including Sam Bond, General Manager of ride service Lyft, and Rukiya Thomas, MARTA Chief of Staff. Curbed Atlanta reported:

Bond noted that currently, one out of every five Lyft trips in Atlanta originates or ends at a MARTA station. MARTA’s Chief of Staff laid out a vision where driverless cars could be timed to pick up transit riders at their origin, delivering them to transit stations to coincide with arriving trains. The feeder system would work thanks to an economy of scale, helping to bring a critical mass of riders to mass transit.

Frankly, to say “transit is dead” is flat out wrong. The debate of transit vs. self-driving cars is misguided–one does not negate the other, but rather they complement one another. Despite a nation-wide decrease in transit ridership from 2014 to 2015, approximately 50 million fewer trips in the first six months of 2015 than the same period in 2014, Atlanta actually saw a 3.1% increase in both bus and rail ridership during the same period, one of the highest in the nation. That’s still 5.3 billion rides taken off the road in first half of 2015 alone.

This argument of autonomous vehicles over transit ignores both the short term needs of the most vulnerable and immobile, as well as the long term needs of the region as a whole. Failing to invest in transit improvements and expansions based on the unproven potential of autonomous vehicles is not just irresponsible as a governing policy in a competitive metropolitan region, but morally bankrupt as a philosophy. For some lawmakers in Georgia, they seem content to let Jesus take the wheel until autonomous vehicles can meet the demand for transit rather than listen to their constituents. If these lawmakers really are interested in giving Georgians more options than just sitting in traffic, then let the people vote and put the myth of the self-driving car cure to rest, for now at least.

Video Sources:

“Filling up seats in cars: The future of driving” Mobility Lab, via

“How Do We Make Rules For Self-Driving Cars?” The Atlantic

“Why Does Traffic Stop For No Reason?” Discovery News

Further listening:

2 thoughts on “The Myth of The Self-Driving Car Cure

  1. I ain’t usly read dis shit but you is right. If politicians be talkin bout computers technologeez “up just around a corner” to save hoes from traffic hours, then why not say virtual realities gon help too? Jus think of them suit an tie wearin mutha fuckas who be “visiting” the office, gettin at meetings and presentations and shit all hyper real virtuality (like dat Hololense). They dreamin fam! My momma ride the MARTA, she ain’t do no über iphone apps none of that shit. An she ain’t takin no robot car to church neither.


  2. In support of the article’s contention that robocars on freeways induce demand, some additional evidence: a) Toyota: induces sprawl. B) Stanford’s Sven Bieker: steals transit riders. C) Fehr & Peers paper: 25% robocar mkt share induces 10% more vehicle miles traveled (VMT). D) California CA Air Resources Board: VMT increases because of rebound effect, sprawl, and mode capture from transit. Details:


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s