I don’t envy the decisions policymakers must make. There is probably only a 50-50 chance that they can get this right as they stare at a crystal ball and predict the future 20 years from now. But a discussion about infrastructure financing and driverless cars is urgently needed to improve the chances of getting this right.

The most obvious discussion that must be had is the tradeoff in building and maintaining roads versus mass transit. There is no question that driverless cars will encourage more road usage. As the costs of traveling falls, cars will travel more frequently and farther. Driverless deliveries become easier and restaurants and supermarkets will likely expand their delivery radiuses.

A countervailing trend is that a significant section of the population — particularly younger — prefer living in cities, to be carless, and to use mass transit. Firms and cities prefer this because it creates a culture of knowledgeable people being physically close to each other, which fosters innovation. The costs of mass transit must be weighed against these harder-to-quantify, large benefits. The discussion thus far points to the need for mass transit in dense areas.

But the discussion about driverless cars quickly becomes more complicated. More driving will require more roads and more frequent maintenance. And the words new and maintenance will take on new meanings. Although cars will communicate with each other, they will communicate directly with embedded technologies in infrastructure.

Placing sensors in white lines and each exit sign will be expensive. And anyone who owns a smartphone knows how quickly these technologies become obsolete and must be replaced. The first 20 years — before the technology stabilizes — will be challenging.

Deferred maintenance — the current default — will not be an option because bumps, potholes, faded lines, and hidden signage will need to be fixed immediately.

The pressures will be felt more acutely in wealthy suburbs, when the first driverless Mercedes that gets into an accident because of failed infrastructure technologies will likely lead to an uproar among residents. Layers of redundancy will need to be built into the embedded technologies.

Current discussions about liability focus on car owners versus car companies only, but cities may well find themselves being liable. Driving will certainly get safer, but it will not be fun for a city’s legal department when a car falls off the side of the road because the city’s embedded technologies failed.

Entirely new types of maintenance crews will be required, from the educated technician to the super-educated technology manager back at DOT. Of course, this will be replicated from city to county to state to national. (If you want another example of how technology creates winners and losers, the salaries of pothole fillers will lag relative to the salaries of technologists).

Local governments can install this infrastructure incrementally, but one can imagine scenarios where local governments compete to be the first to have driverless-technology-ready roads. This will place enormous stress on big cities that need to fund mass transit and communicable roads. The first city that convinces large companies that they have the best driverless technologies and that convinces parents that they will get their kids from home to school and back safely will be a super winner.

The discussions about driverless cars will soon shift fundamentally, if not with the same fanfare, from the car to cities. The safety of driverless cars will become a jurisdictional issue. The elephant in the room is costs to these jurisdictions. Ladies and gentlemen, the era of paying for miles traveled is upon us.

Rayman Mohamed is an associate professor in the College of Liberal Arts and Sciences at Wayne State University.

Read or Share this story: