‘Tis the season for start-ups to make promises about how their self-driving cars will be imminently deployed — and for us to reflect upon the unfulfilled promises of yore. Earlier this month, Zoox, a five-year-old company that makes a four-passenger autonomous vehicle, rolled back into the spotlight with a new concept to solve urban transportation problems. And if cities use this technology correctly (this time), it might actually work.
Self-driving minibuses have been puttering around cities for a few years now: tiny, user-friendly, often ridiculously cute. They’re good solutions for shuttling people through office parks and college campuses, and there are pilot projects in a bunch of places to prove it. The claim, made by Zoox and other start-ups, is that electric, shared autonomous vehicles (known as AVs), when deployed at scale, mean a future of less car ownership and less land set aside for parking, all while more efficiently routing passengers to better consolidate trips (with, for example, less deadheading than Uber or Lyft).
Don’t misunderstand me: A bunch of new electric vehicles are not going to save us. Cities need fewer cars, and people who live in cities need more shared, low-emission options for getting around. There are also a great many autonomy issues to be worked out when it comes to accessibility, privacy, and trust. But let’s focus just on the vehicle itself for now. Unlike the customized Chrysler Pacifica minivans that Waymo is using as robotaxis to move people around Phoenix, Zoox is something completely new. For one thing, the interior is arranged carriage-style, meaning you sit face-to-face with the other passengers, on what looks like extremely well-appointed upholstered seats. That design saves a little energy, the company claims, because the vehicle doesn’t need to turn around — it has no back or front, and it can even slide sideways into a parking space. Zoox also doesn’t have a steering wheel. There is no way for a human driver to override its navigation. That’s actually the goal for safety experts: to achieve what’s called “level 5,” or fully autonomous travel, in vehicles with no opportunity for physical human intervention that are limited to 25 miles per hour or less. It’s easy to envision how summoning one of these might make certain aspects of getting around urban centers easier by shortening a walk home from a train station (known as “last-mile” connections), making it easier to load and unload your groceries, and having the potential to replace specialized access services for people who use wheelchairs or other assistive mobility devices with one vehicle that can serve everyone.
Zoox’s debut is refreshing because most of the headlines about autonomous vehicles are about privately owned self-driving cars developed by auto manufacturers. If we all just go out and buy those, there will be minimal improvement of the climate or congestion, and we may, in fact, be introducing new safety risks. Apple, for example, announced earlier this month that it would be producing a self-driving car for consumers by 2024 (after previously promising such a vehicle by 2021). Earlier this week, Waymo CEO John Krafcik, seemed to confirm that Waymo will never make cars but rather that it’s all about the AI, which will be deployed across a wide variety of vehicles, including ones for private use. And as a fully self-absorbed Elon Musk announced last week, his Teslas will offer a “full self-driving” mode by next year, which is factually untrue — Teslas will still have steering wheels and humans will be able to “take over” driving — and also places a completely unregulated, untested technology in the hands of human operators, some of whom Musk has encouraged to play games while they drive.
As many observers have noted, the federal government has taken a very hands-off approach to which types of passenger vehicles should be allowed to incorporate autonomy (As Trump-appointed USDOT head Elaine Chao infamously said when asked about AVs, “the market will decide.”) Changing that should be a top priority for Transportation secretary appointee Pete Buttigieg, who will be heading up a department where AV regulation has so far been extremely uneven. In February, a woman in Columbus, Ohio, was hurt when the AV minibus she was riding made a sudden stop — and even though it was going just 7 miles per hour, her injuries were serious enough for a hospital stay — and federal agencies banned its operator, Easy Mile, from carrying passengers in the U.S. on its fully autonomous electric shuttles. Federal agencies, however, did not order Uber off the road after its semi-autonomous hybrid Volvo SUV killed a woman in 2018 (although Uber has now sold off that division to a company called Aurora, started by one of Waymo’s original engineers). Zoox, to its credit, welcomes this regulation, saying its vehicle is purpose-built to anticipate such concerns, laden with more LIDAR sensors than it needs, and committed to rigorous testing procedures.
The truth is, a future where Zoox scoops up a couple on a date for the final ten blocks of a journey after a long subway ride, or allows a family of four with a stroller to comfortably travel a short distance together, is still a ways away. Making the case for why such a vehicle needs to be autonomous instead of hiring drivers, and balancing the potential for innovation while equitably fulfilling on-the-ground transportation needs, will remain a major challenge for cities. (Zoox was acquired by Amazon in June as part of its plan to move people around cities along with packages; will it require a Prime membership to board?) But the way to think about Zoox’s design is not as a smaller bus but instead as a shared replacement for a household’s car that performs tasks much better than that car ever would. That includes how Zoox gets to you, which is why car-share services like ZipCar and Car2Go have largely failed in the U.S. — they’re cars, and cars remain a pain to book, store, and navigate in cities. It might plausibly remove a block’s worth of fossil-fuel-powered vehicles from the road and replace them with a single zero-emission vehicle that starts saving all of those car-owning families money right away. (It’s particularly easy to imagine that, say, a lot of two-car households would become one-car-plus-a-Zoox-membership households without much persuading.) Basically it’s what ride-hailing said it would be and never really was — clearly integrated with, priced similarly to, and designed to serve rather than undermine existing public transit.