Robotaxis and the case for public policy accelerationism

Could this disruptive technology force a reckoning with the role of cars in cities?

Robotaxis are here, and we are unprepared, in almost every imaginable sense, to deal with them.

That’s the thesis of my piece in this month’s edition MIT Technology Review. But there’s another dimension to my argument that I didn’t get a chance to flesh out: At a time of rapid, unpredictable technological change, robotaxis highlight the need for a more agile and opportunistic public policy response. Robotaxis just might create new opportunities to advance moonshot urbanist goals by forcing a reckoning with the role of cars on city streets.

I know this sounds counter-intuitive, but bear with me. It’s helpful to start with some context on the current state of the robotaxi industry.

There’s a disconnect between perception and reality when it comes to robotaxis, partly due to all of the wolf-crying of the broader autonomous vehicle industry. For years, autonomous vehicle coverage parroted the industry’s hype machine. But after the hype bubble burst around 2020, the news media and commentariat over-corrected, adopting an aggressively skeptical stance. In particular, media coverage has failed to illuminate the significant differences between personally-owned autonomous vehicles, which remain years away, and robotaxis, which are roaming the streets of several American cities as you read these words. 

There’s still a lot of uncertainty about the future of robotaxis. But I think companies like Cruise and Waymo are advanced and entrenched enough — they could soon be authorized to provide all-day driverless service across all of San Francisco — to merit a shift in the discourse. The most pressing question about robotaxis is no longer whether they work. It’s what we should do about them.  

There are very serious potential consequences to a world crawling with robotaxis. Without adequate safeguards, this technology could induce more automotive travel, increasing congestion in cities and further undermining beleaguered public transit systems. Professional taxi and Uber drivers, already a struggling class, could soon see their jobs automated out of existence.

What’s more, it’s not clear what problem robotaxis will solve. As David Zipper has pointed out, robotaxis in cities like San Francisco replicate well provided taxi, ride hail, and public transit services. Safety benefits will be marginal until robotaxis and other autonomous vehicles make up a significant share of cars on the road. And unlike, say, a new subway line, robotaxis provide no capacity or travel time improvements compared to ordinary cars. The biggest innovation that robotaxis offer is the elimination of labor costs. Some of those savings will be passed on to consumers. But mostly, robotaxis represent a massive victory of capital over labor. Finally, we’ll see the realization of Travis Kalanick’s dream of ditching “the other dude in the car.”

Despite all of these concerns — or, more precisely, because of these concerns — I think robotaxis open a window of opportunity for advancing a more human-centric urbanism, and, potentially, other progressive policy goals.

Call it “public policy accelerationism.” It’s a riff on the Marxian notion that technology and capitalism should be allowed to achieve an advanced state in order to produce the conditions necessary for a proletarian revolution. Public policy accelerationism is much less sexy, but, I think, more realistic and uniquely relevant to the present moment.

The idea is to leverage the novelty, disruption, and visibility of robotaxis to jumpstart fraught policy discussions that will need to be addressed sooner or later. Robotaxis have no existing constituency of users, and will likely be a locus of resentment for many city residents. Unlike Uber and Lyft, which brought some similar problems, robotaxis look conspicuously different from ordinary cars. And they’re irresistible to the media.  

“Regulating robotaxis” would be a politically popular rallying cry that can be used to frame conversations around car-free streets, congestion pricing, vehicle miles traveled taxes, and curb management. Perhaps, robotaxis could be targeted to pilot these sorts of initiatives, allowing the public to get used to them and appreciate their benefits before they’re rolled out to all vehicles.

Congestion pricing fees that disincentivize the use of robotaxis in busy areas at peak times could start to provide a new source of funding for improved or free public transit. Robotaxi-free corridors could be rolled out on major transit or bike routes, priming the public for the notion that certain streets are prioritized for certain modes. The fact that Waymo’s electric Jaguars are exempt from the gas taxes that pay for road maintenance offers a vivid argument in favor of vehicle-miles travelled taxes. And since robotaxis offer such a novel user experience, they could present the perfect moment to introduce requirements limiting pick-ups and drop-offs to one or two designated areas per block.

I know very little about labor policy, and wouldn’t want to speculate as to the best remedies to the issues posed by robotaxis. But the same general principles apply. Robotaxis will provide a dramatic illustration of the automation of work. They could be used as an entrypoint for challenging, necessary conversations about what the automators owe the automated.

There’s a very logical rebuttal to the game of 3-D chess sketched out above: Instead of going through of all this trouble, why not just ban robotaxis? It’s a good question, but the wrong one. For one thing, the history of technological development in the US suggests that’s very unlikely to happen. During my reporting, I came across nothing that would suggest an impending “tough on robotaxis” pivot from regulators. With robotaxi fleets already on the streets in China, federal regulators are probably loathe to slow down this strategic industry. Interstate economic competition (think: Gavin vs. Ron) has helped give robotaxis a warm reception in states that fancy themselves centers of innovation. Secondly, robotaxis, or autonomous vehicles writ large, do have potential upsides in terms of safety and improving access to transportation. Smart regulation, rather than outright bans, could ensure this technology goes where it’s most useful: in places or at times of day with poor public transit and ride-hail service.

To be clear, I’m not holding my breath for an enlightened policy response to the rise of robotaxis. I’m merely suggesting that there is a needle to be threaded, and that with the right combination of activist pressure and political power, robotaxis could serve as the spark for a more human-centric urbanism. For now, the trends are mostly moving in the opposite direction. Robotaxis are already being held up as an argument against improved public transportation in cities like Austin. Robotaxi-friendly street redesigns are next. Michigan is working on developing a dedicated autonomous vehicle corridor between Detroit and Ann Arbor. Even cities that want to regulate robotaxis, like San Francisco, are largely unable to do so because of state preemption.

Which brings me back to public policy accelerationism. Cities that want to mitigate the negative impacts of robotaxis are probably going to have to mitigate the negative impacts of cars. Robotaxis only add urgency, and maybe, political impetus, to the same old set of problems. Regulating robotaxis might be the rallying cry, but regulating cars is the bigger objective.

In response to the arrival of robotaxis, urban planner Jeff Speck acknowledges that cities don’t have much regulatory power. But what cities can do is “legislate with lanes,” he writes in the new edition of Walkable City. “You can allow the growing traffic pressure of a second auto age to ream out your cities anew, or you can decide how much additional space in each street to hand over as demand skyrockets. The proper answer, by the way, is zero.”

I had hoped that a version of public policy accelerationism would attend the emergence of bike and scooter share, and it did to an extent. Most American cities have far better bike infrastructure now than they did a decade ago. Flooding cities with thousands of additional bikes — as in New York, DC, SF and Chicago — or thousands of scooters — as in Austin, Santa Monica, and Nashville — undoubtedly played a role in this ongoing transformation of city streets. All of those two wheelers and their riders had to go somewhere, giving cities a new political and financial (in the form of permitting fees) push for infrastructure. In the grand scheme of things, though, these changes have been relatively modest. The car-centric status quo dies hard.  

Robotaxis are in a different league of disruption. They are, quite literally, an outgrowth of the AI revolution. The changes ushered in by AI could be very fast, very dramatic, and very weird. I hope the public sector, and people who work on public policy, are ready to respond. In this brave new world, it will be valuable to have an affirmative vision of the future we’d like to see, and then opportunistically leverage technological change, or the political fallout from it, to move toward that future.

Urbanists are fortunate to have such a vision. Robotaxis, tunnels-for-Teslas and the like will continue to present new challenges to the future urbanists want to see. But they also provide the opportunity to draw new battle lines, to redefine the terms of a struggle that — let’s be honest — hasn’t been going all that well anyway. Let’s take whatever shot in the arm we can get.