One of the standout promises made by Elon Musk during the Autonomy Day investor event was that Tesla vehicles would be able to operate as “robotaxis” on an autonomous ride-hailing service by the end of 2020. Anyone familiar with the research on the effects of Uber and Lyft on cities should be terrified.
The good thing is that this promise, like so many others made by Musk, is unlikely to come to fruition. He promised autonomy would be ready back in 2016 and that an autonomous Tesla would cross the United States in 2017 — neither actually happened. And even though Musk is promising autonomy will now be ready by the end of 2019, that flies in the face of the timelines put forward by the rest of the industry as Musk rejects the LiDAR sensors and high-definition mapping that experts say will be necessary.
However, Musk’s robotaxi proposal shows what the tech industry aspires to do if autonomy ever reaches a level where it can handle most driving situations, and it’s pretty clear it would be disastrous for cities. A million robotaxis within a year would cripple urban transport systems, but that doesn’t bother Musk because he refuses to accept cities have spatial constraints.
The high urban cost of ride-hailing
A growing body of research has proven that existing ride-hailing services are having negative impacts on cities — the opposite of the alleviated congestion, reduced car ownership, and last-mile transit connections the companies promised. Uber has 3.9 million drivers globally, with an estimated one million in the United States, so the effects it’s had would get far worse if a bunch of new Tesla robotaxis were added to the mix.
Ride-hailing companies add 5,700 vehicles to San Francisco’s streets during peak weekday hours — 6,500 on Fridays — or more than 15 times the number of taxis because there’s no cap on ride-hailing vehicles, according to the County Transportation Authority. Those extra vehicles create 170,000 additional daily trips or 570,000 vehicle miles of travel (VMT) concentrated in the densest and most walkable parts of the city, and that has the effect of making congestion far worse. New York City has a similar problem: ride-hailing services added 600 million VMT from 2013 to 2017, slowing traffic and making bus services less reliable.
They’ve also taken riders from more efficient transport modes, worsening the situation. A study of seven major U.S. metros found that 49 to 61 percent of all ride-hailing trips would have been made by walking, cycling, taking transit, or not made at all had the on-demand services not been an option. Researchers at the University of Kentucky estimated that Uber and Lyft had the effect of reducing rail ridership by 1.29 percent annually and bus ridership by 1.7 percent for every year they were active in 22 U.S. metro areas. That’s a huge hit that’s reflected in declining transit ridership across the country.
These are profoundly negative outcomes that would be compounded by Musk’s robotaxi service, but there are other important reasons why it could be even worse than Uber and Lyft.
Inducing more inefficient vehicle trips
Ride-hailing services spend about 40 percent of their driving time without a passenger. As a result, every mile of personal driving replaced by a private service like UberX adds 2.8 new vehicle miles, while shared services like UberPOOL add 2.6 miles. However, Musk predicted that his robotaxis would be even less efficient with 50 percent of driving time without passengers, contributing even more to congestion.
Further, part of the reason Uber and Lyft are so popular is that they’re cheap, and that’s no accident. Transport consultant Hubert Horan previously estimated that Uber passengers only pay 41 percent of the actual cost of their trip and the company’s recent public filing confirmed it will have to keep subsidizing fares to remain competitive. Those low prices, which cause the company to lose billions every year, are also responsible for attracting customers from transit and other cheaper transport modes. But Musk estimates robotaxis will be much cheaper.
Whether his estimates can be trusted are questionable, but if they’re accurate autonomous Teslas would be far more affordable than existing Uber and Lyft services — $0.18 per mile compared to $2–3 per mile — and would create many more vehicle trips, essentially shutting down our already swamped urban streets.
Musk’s robotaxi vision might seem great to someone who can’t imagine any other way to get around cities than being stuck in a car, but it’s actually really dystopian. Without regulation and taxation, a million robotaxis at $0.18 per mile would cause a boom in automobile use and traffic at the very time that we need to be reorienting our cities away from cars and toward transit, cycling, and walking.
Automobiles are not the future
The good thing to keep in mind is that Musk’s figures are undoubtedly wrong. After claiming Tesla owners could make $30,000 per year by enrolling their vehicle as a robotaxi, he said he “randomly threw some numbers on there,” and a new study out of MIT suggests that autonomous taxi fleets would actually be more expensive than personal vehicle use. Musk has a history of misleading people, and robotaxis are no exception.
However, it does provide even more reason to be vigilant when discussing the what autonomous vehicles could do to cities to ensure the public isn’t being misled about what’s possible. The truth is that the future of urban mobility is transit, cycling, and walking, not autonomobiles driven by humans or artificial intelligence. Self-driving cars will never be a mass mobility option, and people shouldn’t be given false assurances to the contrary.
Yet again, Elon Musk has set out to mislead the public—he’s good at that—and the robotaxis are another in a long string of his lies about transportation. They won’t be as cheap as he claims, nor will they even hit the streets on his timelines—if ever. We need to stop listening to tech billionaires whose visions of the future aren’t in the best interests of the majority of people and retake control of the narrative. The future of urban mobility is exciting, but it’s not autonomous.