Expectations, uncertainties, and policy choices for self-driving cities

A guide to where things stand on this most promising — but highly uncertain — technology.

By Rohit T. Aggarwala

A driverless mini-bus

A driverless mini-bus serves passengers on a medical campus in Berlin, Germany, in March 2018.
(Kay Nietfeld/picture-alliance/dpa/AP Images)

If we know one thing about how technology is changing urban mobility, we know that autonomous vehicles are going to be a revolutionary force.

If we know two things, however, it is that we don’t know much beyond that about AVs. How they’ll be adopted, what affects they’ll have on cities, when they’ll arrive, what we should be doing about them — all of these outcomes are highly uncertain. This uncertainty often leads to irrational optimism, unnecessary pessimism, or a misplaced focus on things that aren’t as important as they might seem.

To offer a measured introduction to where things stand, I’ve published a paper entitled “Autonomous Vehicles and Cities: Expectations, Uncertainties, and Policy Choices,” which grew out of a memo I presented this past February at a roundtable at the Council on Foreign Relations. The proceedings and papers of that event were published in a volume called Digital Decarbonization: Promoting Digital Innovations to Advance Clean Energy Systems, edited by the organizer, Varun Sivaram. It’s available as an ebook or as a pdf download.

My conclusions fall into three categories:

Reasonable Expectations

  • AVs will be on the roads of many cities over the next five-to-10 years.
  • AV fleets will enable cities to reshape the urban street and reduce the need for on-street parking, but they will still need maintenance and staging areas.
  • AVs will compete with low-frequency and low-speed public transit but will not substitute for high-capacity, high-quality subway and bus systems.
  • AVs will reduce crashes but maybe not by the 90 percent figure many have promised.
  • Not all AVs will be electric from Day One.

Key Uncertainties

  • The employment impact of AVs will depend on whether traditional driving jobs lost are replaced by the unclear, but likely to increase, jobs required for maintenance and customer service.
  • The traffic congestion impact of AVs will depend on whether new driving encouraged by the lack of a driver (such as repositioning vehicles, more frequent deliveries, etc) are offset by an increase in shared rides.
  • Most companies working on AVs today are espousing a for-hire, fleet approach to AV deployment, but AVs will clearly be available for private ownership. Whether people continue to own their own cars or rely on for-hire vehicles will depend on marketing, habits, and policy.

Policy Choices

  • Safety standards will have to be set by government and will make tremendous differences in the overall impact and public acceptance of AVs.
  • Some of the benefits of AVs — such as the elimination of on-street parking — will require a decision by local government to redesign the streets. Waiting for 50–60 percent adoption of AVs will likely reinforce existing street usage patterns.
  • Road pricing may be the most important tool government has to ensure that AVs do not lead to overconsumption of road space and to promote the shared usage of for-hire AVs. Whether and when to implement pricing, and how, is a critical policy choice.

Of course, if we know a third thing about technology and urban mobility, we also know this field is changing every day. Since I wrote that paper earlier this year, Waymo announced a partnership with Jaguar on electric propulsion; SoftBank invested $2.25 billion in GM; and Tesla’s autopilot was involved in another fatality. In addition, Dan Sperling published his excellent book on AV issues, Three Revolutions, and the World Economic Forum and the Boston Consulting Group published a paper on their AV work in Boston, which offers further indication that the impacts of AVs are likely to be highly nuanced.

These are the types of trends and open questions we’re exploring as we develop plans for Sidewalk Toronto, with an eye toward helping urban neighborhoods capture the benefits of autonomous vehicles. That goal leads us to several areas of focus.

One is how AVs can help redesign city streets for pedestrians. Most of the discussion of AVs has focused on people inside the vehicles: travel time reductions, safety improvements, cost savings, adoption rates. We’re thinking about how AVs enable streets with more room for broad sidewalks, linear parks, shareable spaces, and indoor/outdoor retailing. That focus means thinking about the safety standards needed for AVs and pedestrians to share a dense setting. If we train our vehicles to maximize speed like so many human drivers do, we won’t create streets that are more welcoming to people.

A related area we are thinking about is what rules and systems are needed to ensure that people use AVs in a way that’s good for the city around them. A pricing regime that encourages shared rides seems necessary — WEF and BCG concluded similarly — but we should do that in a way that avoids the coarseness of traditional cordon pricing or tolling systems.

We will have a lot more to say about the role of AVs in cities in the months to come. But as we look ahead, if there’s one more thing we can be certain of, it’s that our expectations for AVs will have to evolve as rapidly as this most promising of technologies.