A couple of years ago, John Niles and I started writing, blogging, and speaking about getting ready for autonomous vehicles. We called this work, collectively, The End Of Driving in recognition of our certainty that technology would displace the driver just as the automobile had replaced the horse. We’re still certain of this.

Then, as now, we were convinced that no one can forecast how this will unfold in any useful detail—i.e., useful enough to write infrastructure plans or social policy. We took to describing the next quarter century as the most difficult transportation planning horizon ever faced by urban planners.

The prognostication ranged then, as it does still, from no more personal automobile ownership to no more public transit to instant-and-perfect robo-cab within 2 minutes of a request, to a clean and perfect car for every garage. And with these would come green utopias of no accidents, no parking, no congestion, and info-cocooning on the way to work. Or maybe dystopias of sprawl, transportation inequity, increased congestion, and robots chauffeuring parent-free five-year-olds to ballet.

In the March 7^th issue of Time magazine, Xerox Executive David Cummins, thinking about all the immediate, incremental changes, comes off as indifferent. “Cars parallel park themselves now. Cars speed up and slow down on their own already. Cars have all kinds of accident-avoidance technology. And you are going to have more and more and more of that introduced over the next three to five years. By the time that first car rolls off the factory line without a steering wheel, it’s not going to be that much of a shock. The collective response may be more of a shrug. As in, ‘It’s about time.’”

From the beginning, at EndOfDriving, we’ve thought about what kinds of social directions this change could take—or should take. We thought about invented futures rather than wait-and-see futures. We understood that the world is trending from one billion motorized vehicles in 2010 toward four billion by 2050, but we wondered if the implied quadrupling of motorized trips or kilometers to be demanded by 2050 could be provided—with the requisite sharing—by only one billion vehicles.

We imagined how magical it would be always to have a vehicle arrive on command and whisk us to wherever. And we wondered whether we could understand and overcome all the operational and social barriers that were implied. We noted that the fast-rising TNC world of Uber and Lyft was not only wonderfully disruptive of the taxis many disdain, but it was also ushering in an ugly disruption: TNC’s cherry-pick the best taxi customers—young, well-groomed digerati—leaving poorer and disabled patrons as cast-offs for the now dying taxi industry.

We understood that driverless robo-fleets could erode the need for household ownership, but also for the city bus and that this would force rollbacks on massive municipal pension obligations if transit systems were to collapse.

We calculated that since vehicle populations take 20 years to double, and fleets take 15 years to turnover and that if error-free and affordable robotics were available optimistically in 10 years, it would take 40 years until the last driver was retired—by which time, of course, we’d have 4 billion extant vehicles to counter the promised advantages of robotic vehicles in the first place.

Since then the hype, both positive and negative, has grown in volume and confusion. Academic experts caution that reliable, full autonomy is at least 20 years away and pervasive, driver-out, autonomy, more like 40 or 50. Still lighter-headed commentators contemplate radical autonomous scenarios starting in 2020 and becoming pervasive in the following ten years, perhaps encouraged by the optimistic projections of automotive marketers. Thinking about nuanced social changes happens less often.

We are not questioning the ability of automotive technologists or even AI practitioners; rather we are questioning urban and social readiness for many dozens of overlapping, concurrent and contradictory changes.

When writers weigh multiple aspects concurrently, the contradictions mount quickly. David Ticoll, University of Toronto, sees AVs mitigating congestion if combined with mass transit while at the same time AVs stimulate more use of cars. How will that push-pull combination work out?

…factors that cause congestion – capacity bottlenecks, traffic incidents, poor signal timing, weather, construction, and uncertain travel times – can be reduced or mitigated by vehicle automation, especially when combined with public mass transit. At the same time, the accessibility and convenience of AVs is likely to stimulate greater use of cars and minibusses. [pg5]

As of 2016, there is no consensus whether the autonomous vehicle will enable better transit or no transit. There are some proponents for each forecast. But for most pundits the future transit world seems to be destined to be crushed no matter which of the most popular two scenarios prevail: pervasive robocabs or a renewed juggernaut of private household vehicles.

In any case, we are clear that transit will be disrupted. We are tentative on whether it will be disrupted in a way that lets transit managers evolve services and transform quickly to more effective formats. We are fearful that transit authorities will abdicate completely to the new private vehicles of the future and to the robo-fleets of the rising TNCs.

Urban leaders need to start preparing now for the decisions they will soon face.

Bern Grush, Toronto, 2016