Elon Musk wants to reinvent the ride-sharing industry, the Tesla way. As the fleet’s autonomous driving capabilities improve, Musk is proposing that customers could put their cars to work while they’re not using them.
During a “Tesla Autonomy Day” event this week, Musk gave the clearest picture yet of an idea he’s been teasing out for months. Tesla owners will soon able to loan out their self-driving cars to schlep people around, competing directing with human-powered ride-sharing services like Uber and Lyft. It would be called the Tesla Network. Musk hopes to have 1 million robo taxi-ready cars on the road by 2020.
It sounds pretty great. Car ownership remains a necessity for many Americans, as 85 percent use a car to get to work. They’re increasingly taking out risky, high-cost loans they can’t afford in order to buy them. Subsidizing car ownership using new software, then, seems like a win-win, but transit researchers and urban planners aren’t so sure.
For one, how are we going to fit all these roving robo taxis onto already crowded streets?
Autonomous Driving Poses an Urban Planning Problem
The main concern of transit advocates? A convoy of self-driving cars like the one envisioned by Musk will create more traffic, have detrimental environmental effects, and detract from shared means of transportation. This scenario was according to three transportation experts who shared their concerns about Musk’s plan with Inverse in a series of interviews this week.
“This is likely to lead to much more traffic, taking trips away from public transportation, a lot more congestion, and more environmental impacts because the [manufacturing of batteries] affects the environment even if it’s better than gasoline cars,” Adam Millard-Ball, a professor of environmental studies and economics at the University of California at Santa Cruz, tells Inverse.
The main problem, Millard-Ball says, is that congestion in some major cities is already bad, enough so that any uptick in congestion at all will make them vulnerable to gridlock. This is also exactly where Musk envisions the Tesla Network being most popular: urban cores.
American drivers spent some 97 hours in traffic over the course of 2018, according to transportation consulting firm INRIX, an opportunity cost of some $87 billion.
The cost of traffic in the top 25 most congested US cities topped $400 million per city, and the fastest speed for the “inner city last mile” is about 17 mph. (Hello, Dallas and Phoenix!) Municipal planning experts say inserting self-driving cars into an already congested environment will make traffic worse because they won’t have to park and will stay on the roads, in traffic, for longer.
In addition to simply not having to park, autonomous vehicles will be incentivized to cruise around city blocks to avoid parking fees or entrance payments for a parking garage. This could exacerbate gridlock issues in urban areas even if only a couple thousand autonomous cars are introduced.
In a February reserach paper published in Transport Policy, Millard-Ball generated a simulation that factored in just 2,000 autonomous cars in San Francisco, and his findings were not reassuring. The AVs slowed down traffic to less than two miles per hour, which understandably would bring anyone to the brink of insanity. Many of these cars could even be empty as they cruised around waiting for the next fare, like taxi cab drivers sometimes do today.
Empty Teslas — or any other self-driving car used as a robo taxi — taking up the road was also a concern raised by Sarah Kaufman, associate director of New York University’s Rudin Center for Transportation, who tells Inverse that these networks will eventually lead to a hollow kind of gridlock. Large segments of bumper-to-bumper traffic will be caused by cars that would normally be off on the street in a parking spot.
“What full [vehicle] autonomy will do is allow us to eliminate parking,” she said. “Many of these roaming cars won’t even have passengers in them. If we have zero-passenger vehicles cruising the streets, we could reach a point where there is just endless gridlock that city streets simply cannot absorb more vehicles, empty or not.”
The “Induced Demand” Problem
The potential for these near-future traffic jam scenarios (fun!) boil down to the fact that ride-hailing services have made traveling in cars cheaper, easier, and more appealing.
This “induced demand” problem obviously didn’t start with Tesla. Companies like Uber and Lyft have long worked to make cars available at the touch of a button. It’s a problem of which Musk is aware. He even referenced the theory of “induced demand” when talking about The Boring Company in December. While an additional lane on the highway induces more demand as to create only a bigger traffic jam, Musk argues that a limitless number of underground tunnels could beat induced demand once and for all.
By pursuing autonomous driving technology, companies like Ford, Waymo, and Tesla all run the risk of making the induced demand problem even worse by eliminating the need to have and pay drivers entirely.
Induced demand from ride-sharing has already been documented in some studies. A 2017 paper from the University of California at Irvine’s Institute for Transportation Studies estimates that between 49 percent and 61 percent of rideshares either wouldn’t have been taken at all or would have been taken using bike or transit.
How to Keep These Sweet Robo Taxis From Screwing Up Our Roads
The influential transit consultant and Human Transit author Jarret Walker tells Inverse that cities need to begin developing new policies to accommodate all these new cars.
“Induced demand is a well-documented effect of Uber and Lyft, and it will be worse if cars are autonomous,” he said. “The only solution in dense cities is the fair pricing of road space.”
Translation: congestion pricing, the sometimes controversial policy of charging drivers more when they enter certain parts of the city. Walker argues that by eliminating parking fees and with fewer public transit customers, congestion pricing will be the best way for cities to both manage congestion — by encouraging people to park and hop on a train or bike — and profit off of it, by raising revenue from the fees to improve public services. The plan would be to treat the most frequently used roads as prime real estate that drivers (or in this future case, autonomous car owners) need to pay for if they want to use it.
These policies have already been enacted with great effect. Drivers in London pay a fee of about $15 to enter the city center and cities, and between 2002 and 2014, the number of personal use cars in the private zone fell 39 percent. Large, walkable, and sparkly clean cities like Singapore and Stockholm, Sweden, also use similar rules. New York City is also close to approving a similar policy that is projected to raise about $1 billion annually from drivers traveling to the most congested parts of the metropolis.
“It’s a very effective policy to reduce congestion and to raise money to improve [public] transportation systems,” explains Millard-Ball. “Right now, we give roads away for free, incredibly valuable urban real estate that we give away to drivers. That’s why we have congestion, because it’s free to use.”
Fortunately, there’s still time for urban planners to prepare. The timeline for autonomous cars taking to the streets has slowed down. Accidents, crashes, and even some deaths caused by prototype vehicles have forced companies to step back and assure the public they are taking safety issues seriously. But that doesn’t mean we won’t eventually have to reckon with the sweeping changes AVs will bring to our roads.
Tesla, for one, is set on making autonomous vehicles a reality. The opportunity is too great. After all, it’s an innovation that could single-handedly eliminate all the accidents caused by drunk drivers and texting and driving and save thousands of lives each year. But cities and countries need to begin future-proofing their streets to avoid what could be endless, robot-induced traffic jams.