Detroit drives future

Motor City can lead on real-world autonomous vehicles

San Jose, Austin, Cambridge, San Francisco … Detroit?

Most people don’t think of the Motor City as an incubator of tech innovation. But as both Big Tech and nimble startup culture have turned their attention to the auto industry over the past half-decade, lured by the promise of driverless cars, both sides have seized on a new opportunity for partnership — one that’s partly the product of a joint moment of necessary humility.

As the Silicon Valley mantra of “move fast and break things” meets the unforgiving reality of the urban road, Detroit is becoming the natural home to dozens of auto-tech ventures. Harnessed correctly by rational regulation, Detroit can view the emerging industry of autonomous-vehicle technology not as just a job creator, but as a way to improve quality of life for residents and commuters.

A decade ago, as Detroit was hurtling toward municipal bankruptcy and the region’s auto industry was mired in economic crisis, it might have seemed obvious that the next generation of automotive research and development would find its natural home in Silicon Valley. On the software side, Uber and Lyft were building headquarters in San Francisco. On the hardware side, Tesla, too, was making California its home. The West Coast would further build on its advantage: a critical combination of tech brainpower and startup brashness. Who needs lumbering old car companies?

But in the past few years, tech has absorbed two lessons. First, building a car is hard. Of course, building Google was hard, too. But a company that wants to build a car from scratch or sell a driverless-tech innovation to an existing auto company must understand how a physical assembly line works, something that the non-hierarchical tech world hasn’t done on a mass scale, but that the Detroit region has long been perfecting. 

Second, it involves an entirely different testing culture — one in which caution and incrementalism are just as important as being first. Uber learned this the hard way last year, when one of its driverless SUVs ran over and killed a pedestrian in Arizona. The Uber vehicle was travelling at high speed on a complex road with only an improperly trained attendant to monitor its path. The attendant predictably got distracted, helping to precipitate the crash.

Contrast this experience to how Ford-partner Argo is testing driverless vehicles — passengerless, for now — in downtown Detroit. The company requires two trained workers in each vehicle. One watches the road, while the other watches a computer; both talk to each other to ensure that the road matches the computer map and vice versa. As the cars learn, the company keeps the vehicles confined to limited, mapped areas, being sure not to push it further than it can go. 

May Mobility, an Ann Arbor startup, takes the same approach to its downtown Detroit shuttle, which ferries passengers from parking lots to office buildings in partnership with office-property manager Bedrock. The shuttle, confined to a mapped route and attended by a staffer, travels at slow speeds, and will always err on the side of caution. It stops at yellow lights, for example, and stops if it has any question about a pedestrian’s path, letting the attendant take over. 

In terms of how to test and build a complex vehicle or a component for such a vehicle in real-world conditions, it’s Detroit, not Silicon Valley or Boston, that has the advantage of a dense network of existing brainpower. Detroit is home to 17 “original-equipment manufacturers,” or OEMs, in the industry’s parlance, and 96 of the nation’s top 100 suppliers, as well as three research universities in the region adept at turning out grads for the industry. It also has an infrastructure of test tracks and regulatory relationships. 

At the Detroit Regional Chamber and Michigan Economic Development Corporation’s PlanetM landing zone, companies ranging from Lyft to the pedestrian-safety startup Derq have chosen to take advantage of subsidized downtown office space for the chance to collaborate with competitors and potential partners and customers. And though much of the country boasts diverse weather, it’s not a bad thing that Detroit offers rain and snow as well as sun for testing AV resilience.

The city of Detroit, too, can benefit from a thoughtful approach to driverless tech, particularly as the wild predictions of a few years ago — that everyone would be riding around in a driverless car by the year 2020 or so — have dissipated. The technology to make a vehicle fully autonomous — that is, a car that drivers itself — is complex, difficult to power, and difficult to maintain, meaning that it’s likely to remain confined to carefully supervised shuttle services and the like for a while.

That doesn’t mean, though, that the city can’t benefit from incremental improvements now as it considers how AV fits into its broader transportation strategy for the future. On this front, Detroit can benefit from some existing disadvantages. On providing mass transit, it has nowhere to go but up. Fewer than 8% of Detroit workers take mass transit to work, according to the U.S. Census.

That means that the city doesn’t face the same problem faced by cities such as New York and Boston: that using AV to enable shared rides, such as the May shuttle, might simply take people off a more efficient subway system, defeating the purpose. Then, too, far too much of downtown is covered by parking lots or garages, an inefficient use of urban space. Encouraging more mass transit means that as Detroit rebuilds its population, it could transform some of that currently needed parking space into office buildings or homes.

Detroit, then, can lead the way in using tech for mass transit. The city could partner with a company such as Via, which provides shared rides within cities ranging from Berlin to New York, to offer a more frequent and faster bus to and from the airport at a slightly higher price than the $2 bus fare. The city could also work with disabled residents for door-to-door service, combining the on-demand aspect of existing ride-hail technology to map efficient routes for frequent, subsidized, shared vans.

Detroit can also harness autonomous technology to improve safety. Detroit is one of the nation’s most dangerous cities for pedestrians — in part because there are so few people crossing wide-open roads on foot. The best remedy for cutting road deaths is density; walkers and cyclists automatically slow drivers. The next-best cure is design; Detroit is already building protected bike lanes and wider sidewalks in a few key areas, and red light and speed cameras would help, too. 

But tech, too, can play a role. More driverless shuttles on the road could themselves change other drivers’ behavior, calming traffic both proactively, with their slow speeds and turns, and reactively, with their eyes-on-the-street vigilance making it harder for a hit-and-run driver to escape without notice.

Over the past two decades, Americans have adjusted to everyday tech, for better and for worse — and over the coming decades, a well-focused Detroit can be the proving ground to help them adjust to tech on urban streets, for better rather than worse. 

Nicole Gelinas is a contributing editor to the Manhattan Institute’s City Journal, where she is doing a research project on autonomous technology and Detroit, as part of the Manhattan Institute’s Urban Policy Series. Twitter: @nicolegelinas