Announcing The End of Driving

The self-driving future we get depends on the choices we make today

It is February 2025, and today hundreds of people in Phoenix, San Francisco, Los Angeles, and Austin, not to mention Beijing, Shenzhen, and Shanghai, will traverse their cities in a car without a driver.

This is a moment to celebrate. We’ve been working towards this moment since 2005, when the second running of the US DARPA Grand Challenge kicked off the race to driving automation. But we’ve been dreaming of it since the 1939 Futurama exhibit at New York World’s Fair; arguably, since the Arabian Nights’ fantasies of flying carpets.¹

This is also a moment of humility. Also since 2005, we’ve constantly imagined a future where such vehicles are commonplace, and their ubiquity changes everything, for good and for ill. That enthusiasm was misplaced. It’s clear that widespread availability for these vehicles, in wealthy cities at any rate, is still years away.

Years, that is. Not decades. Years.

And that means that now is the time to prepare.

My new book The End of Driving, second edition, co-written with John Niles and Bern Grush, will be published in August. It is available for pre-order here.²

In The End of Driving, we examine the critical choice that faces us:

• Do nothing. Automated vehicles will reinforce our current system of widespread private car ownership; providing new benefits, but also exacerbating old ills

• Act. Specifically, act to ensure that automated vehicles give people choices they do not have today; in so doing, make urban and suburban mobility more equitable, more reliable, and less expensive

When Bern and John wrote the first edition of the book, published in 2018, they were voices of caution, drowned out by a cacophony of hype. They wrote in an environment where many predicted that by today, driving automation would be everywhere, and human-driven vehicles would be as rare as horse-drawn carriages are. Those credulous and outlandish predictions have not come to pass; today, across the globe, there are not hundreds of millions of wholly-automated vehicles, but only a few thousand.

John and Bern were right then to recognize that the self-driving hype wave ignored the technological challenges facing driving automation. They also anticipated our collective failure to grapple with the social, regulatory and behavioral challenges that also face it. To their credit, they admit now that they failed then to appreciate just how difficult the transition would be; they listed barriers (“not a very hard task”, they acknowledge), but often without understanding them. In their defense, understanding them in advance of their emergence would have been nigh-impossible.

But today, those barriers have emerged, and we can understand them. That means that we can have a more grounded discussion about automated vehicles' role in urban transportation.

The core opportunity remains compelling: robotaxis and automated shuttles in common-carrier public service could dramatically reduce the need for private vehicle ownership. Making it easier for people to live without, or with fewer, cars would reduce people’s expenses while increasing their freedom, and would encourage widespread and positive changes to land use, commerce, and—especially—mobility access for those unable to drive or afford cars. This technology, properly governed and deployed, can help create more livable and equitable cities.

Crucially, those benefits won’t emerge automatically from the technology itself. Achieving them will require deliberate policy choices and careful system design. It will require thoughtful preparation.

That is the purpose of The End of Driving: to describe what must be done to achieve the future we want.

Two Diverging Paths

Automated driving stands at an inflection point; what happens in the next few years will shape our cities and communities for decades to come. The End of Driving argues that we stand at a juncture between two futures.

In one, automation primarily enhances privately-owned vehicles. Today, advanced driver-assistance systems (ADAS) makes the driving task easier: adaptive cruise control, lane-keeping, automatic braking, and similar features. Over time, these have become more sophisticated and more common, and will continue to do so, culminating in personally-owned vehicles that, while driving along limited-access highways at least, will not require human input at all. This future is already emerging, as automakers introduce more ADAS features at all price points. It is a natural expression of existing business models.

The result of this development will be that private car ownership becomes more attractive and convenient, which in turn will tend to worsen congestion, sprawl, inequity, and the inefficient use of urban space. Even so, this path leads to a better world than we have now, in that there are some incremental improvements in safety and convenience; but otherwise, the world remains on the same trajectory it is today, missing out on the transformative potential of automation.

In the other future, the nascent fleets of Waymo and Zoox robotaxis we see today grow; supplement themselves with robobuses, the offspring of today’s automated shuttles; and integrate themselves into urban public transit. Today, many bus routes operate in low-density settings and require significant subsidy per passenger; robotaxis and automated shuttles could, if we permit them, replace these routes at lower cost, while making riders happier. In this version of the future, vehicle automation would become a core part of our transportation systems, with a variety of positive results. It would reduce the need for private car ownership, while increasing mobility access for the people who cannot, or choose not, to drive. Further, there would be dramatic reductions in parking demand, improved access for disadvantaged populations, and more efficient use of urban space.

Note that I didn’t say we stand at a juncture between two options, but instead two futures. The first future isn’t something we need to choose; it is the default. It is what will happen if we do nothing. The second future is a choice, because it won't emerge automatically. Achieving it will require us to rethink everything from urban curb design, to the scope of transit operations beyond traditional fixed-route service, to land-use patterns. The book examines this juncture in detail, to show why the first future is our default path, and what specific steps we can take to realize the second.

How to Proceed

Given these realities, what concrete steps can we take today to guide the emergence of driving automation toward better outcomes? In The End of Driving, we suggest three critical areas for intervention:

Firstly, we need appropriate regulation that balances innovation with public benefit. This means moving beyond simple safety requirements to tackle thornier questions: How do we ensure robotaxis serve all neighborhoods, not just wealthy ones? What role should public agencies play in setting service standards? How do we maintain healthy competition while preventing a race to the bottom in service quality? These frameworks must be flexible enough to accommodate rapid technological change while remaining firmly focused on public good.

Secondly, cities need to proactively plan for automated mobility through changes to their infrastructure and zoning policy. Parking spots converted to pickup zones; traffic signals upgraded for vehicle communication; by-laws changed to reduce the minimum number of parking spaces required in new buildings… Many aspects of urban management will require attention and change. Each decision on these matters will either reinforce car dependency or support a transition to shared mobility.

Thirdly, and most importantly, public transit agencies must evolve from operators of vehicles to managers of mobility. Rather than viewing robotaxis as competition, they should integrate them into a comprehensive system that matches each travel need with the most appropriate service. This means using automation to replace low-ridership fixed-route buses with flexible on-demand vehicles, extending service hours and coverage areas, and providing better options for people with mobility challenges. Most importantly, it means ensuring that new mobility services complement rather than cannibalize the high-capacity transit lines that work cost-effectively to serve many passengers at once.

These interventions are interconnected. Success in any one area depends on progress in the others. Moreover, they require coordination across multiple stakeholders: technologists need to understand policy constraints, policymakers need to understand technological capabilities, and everyone needs to understand how human behavior shapes adoption of new services.

Topics We’ll Explore Soon

Speaking of these groups of readers: for technologists, The End of Driving provides crucial context about how innovation—both the tools we have today and the ones we can reliably foresee—will interact with complex urban systems and human behavior. For planners and policymakers, it offers practical frameworks for managing this transition, including specific approaches to regulation, infrastructure, and service design. For transit operators and politicians, it shows how automation could either threaten or revolutionize their services, depending on how they adapt. And for anyone interested in the future of cities and transportation, it presents a clear-eyed look at both the opportunities and challenges ahead. The tools and approaches in the book can help guide this complex transition toward better outcomes for all.

Over the coming months, I will occasionally use Changing Lanes to explore these themes, and more besides. Each of these associated issues of Changing Lanes will examine a critical aspect of automated mobility. Matters we’ll consider include:

• The human side of automation; why some people will remain attached to private car ownership, and how to give them tools that might change that preference

• Mixed traffic: the extended period in which automated and human-driven vehicles will (and must) coexist

• How our cities will adapt, from parking to pickup zones

• The evolution of traffic-management systems

• The role of vehicle-to-infrastructure communication

And most importantly, the several ways we can integrate automated vehicles with our existing transit systems; a theme I have examined before.

Most importantly, we'll keep returning to our central argument: while automation offers tremendous potential for improving urban mobility, realizing these benefits requires us to recognize the choice before us; to plan how we will seize the benefits that thoughtful automation can provide; and to follow through, by making the hard political and policy choices to make it happen.

The end of driving is coming. What the end of driving means is up to us.

1. If you’re wondering, yes, that’s what the TV show was named for.

2. Please ignore the overwritten abstract; the publisher is working with Amazon to change it.