clock menu more-arrow no yes mobile

Filed under:

Lyft’s 10-Year Driverless Car Plan Is a Pipe Dream

We applaud the ambition, but reality will be a tough road to travel

Getty Images
Getty Images

When will our streets be overrun with driverless cars? It remains an open question, with everyone from Audi to Uber investing heavily in robocar technology. On September 18, ridesharing service Lyft planted a flag in the sand. In a sweeping 10-year roadmap, cofounder John Zimmer proclaimed that private car ownership will “all but end” in major U.S. cities in 2025, thanks in large part to autonomous vehicles.

That seems … ambitious. The majority of households in almost every major American city had a car in 2012, according to a University of Michigan study. The same study found that even in New York City, which already has a sprawling, 24-hour public transit system and a massive network of taxis, more than 40 percent of households had a vehicle. Turning any of these places into a driverless oasis will require untangling a knotty thicket of challenges that include technology, infrastructure, regulatory oversight, and consumer skepticism.


Lyft plans to begin shuttling customers in self-driving cars via a partnership with General Motors within the next year, and Zimmer pledges that more than half of the company’s rides will be autonomous by 2021. Here in 2016, Uber is the only company using driverless vehicles with customers in the United States, in a newly launched program in Pittsburgh. Ford has said its mass-produced driverless car, which will be used in a ridesharing service such as Lyft, won’t be ready until 2021. Google, long the leader in advancing driverless-car technology, has no public timetable for when (or if) its cars will be available to consumers.

Point being, we’re still very much in the testing phase of driverless cars. The only places they’ve been used by actual customers en masse are not city streets, but major highways where traffic is more predictable and there are fewer obstacles like pedestrians or cyclists. But even this use case has already proven faulty. Tesla has faced harsh criticism since a driver using the electric automaker’s semiautonomous technology, Autopilot, suffered a fatal collision. Every such crash — and there will be more, inevitably — will make it harder for tech and auto companies to get legislators and customers on board with their nascent transportation revolution.


Making a city amenable to driverless cars isn’t as expensive as building a train system, but it still requires additional investment in America’s shoddy transportation infrastructure. Faded lane markings on old streets would have to be better maintained because they befuddle the onboard cameras that driverless cars use to navigate. Potholes, which the vehicles sometimes mistake for shadows, would be a bigger threat to passenger safety. If a city attempted to eliminate traffic lights, as MIT researchers have proposed, it would have to develop a centralized network that functions as a kind of air traffic control for cars. Columbus, Ohio, is developing an expansive traffic-monitoring system of its own right now. It costs $76 million. What other cities are going to be willing to shell out for the still-unproven tech?

There are other challenges that even money can’t solve. Driverless cars will be harder to roll out in cities with frequent inclement weather because they struggle in rain and snow. Bridges and hills that obscure the view of the car’s camera also present challenges. Tech companies aren’t shying away from these issues — the unusual street layout and harsh winters were one reason Uber picked Pittsburgh — but they’ll delay widespread rollout of the tech. “There are a lot of those types of idiosyncrasies that do represent challenges for the technology that enables driverless cars,” says Jeremy Carlson, a senior autonomous driving analyst at global information company IHS.


The federal government is taking a relatively hands-off approach to autonomous cars. The set of guidelines issued last week by the Department of Transportation do more to address the safety of individual vehicles than how to effectively deploy them in a major city. Companies will have to travel state to state and town to town to hash out the rules of the road, which are likely to vary widely. These municipal battles are likely to take a political bent in some places, as cities experiment with using driverless cars to replace traditional public transit. The same way disruptive companies like Airbnb are forced to fight pitched battles around the country to gain acceptance for their businesses, driverless car companies will find themselves in many statehouses and mayor’s offices.

Consumer Acceptance

Finally: Even if the tech works, the cities pay for the necessary infrastructure, and the legislators deem them legal, driverless car companies will still have to pass muster with a skeptical public. Seventy-five percent of drivers fear riding in a self-driving vehicle, according to a survey by AAA. Despite constant talk about how millennials hate driving and other totally annoying adult things, the vast majority of people still own cars. How quickly can their minds be changed? That will depend partly on price and how much cheaper companies like Lyft can make the use of robotaxis compared to driving yourself around. But their ability to offer inexpensive rides while also building a profitable business depends on scale. Can these companies persuade enough people to hop in a driverless vehicle so that they can drive down costs (read: paid human drivers) and fares? It’s an open question, but it will require a massive shift in the psyche of the average car-owning American.

Before any city truly declares open season on driverless cars, expect to see advocates try to carve out test areas for the tech. A proposal cowritten by a venture capitalist in Seattle and a former Microsoft exec, for instance, calls for a ban of human drivers from a 150-mile segment of the interstate at certain times between the city and Vancouver to make room for robocars. But those kinds of tests will take time. Now at least, thanks to Lyft’s bold proposal, we have a shot clock to keep an eye on.