Late last year, New York Governor Andrew Cuomo announced that Cruise Automation — the driverless car brand of General Motors — would soon begin testing autonomous vehicles on Manhattan streets. The announcement came a few weeks after Governor Cuomo accepted a $17,500 donation from the automaker, a number that would eventually balloon to $28,000 in donations.
About five months after Governor Cuomo’s announcement, a driverless car being tested by Uber in Tempe, Arizona, ran over and killed a woman walking a bicycle across the street. As of press time, the crash appears to have inspired General Motors to put its autonomous vehicle testing plans on hold in New York City.
These two discordant stories — the killing of a pedestrian by a driverless car, and the easy money it took for driverless cars to be potentially let loose in the most pedestrian-dense environment in the United States — spark a few critical questions: How would New York City have to change to accommodate autonomous vehicles? Is that what we want? Do we even have a say?
How would New York City have to change to accommodate autonomous vehicles? Is that what we want? Do we even have a say?
To figure out the answer, we asked Sam Schwartz, New York City traffic engineer and author of No One at the Wheel: Driverless Cars and the Road of the Future (out in November 2018 from Public Affairs), to paint the picture.
“If done right,” Schwartz says, “driverless cars could bring a reduction in congestion, crashes, injuries, and deaths. We could see elderly, disabled, and low-income people well-served by transit, more sharing, a solution for the last mile, and a reduction in parking demand. But there is also a fair probability that vehicle miles traveled will soar, congestion will increase, and a lot of jobs will disappear. Driverless cars could take people out of subways, and slow buses even more than they already are.”
“In the city of the future, the pedestrian has to be the king and queen. But I fear the autonomous vehicle industry, through its unprecedented lobbying power, will suggest that we confine pedestrians behind fences and corrals along the sidewalk, with cattle gates to allow people to cross or not. Some researchers are saying that don’t think they can solve ‘the pedestrian problem,’ so people will have to wear badges with a transponder to tell the driverless car that a person is walking, or biking, nearby.”
The future Schwartz describes — of pedestrian fences and people on bikes legally required to carry transponders — is dystopic but not hyperbolic. As of right now, driverless car technology struggles to make decisions around “randomness,” like children, wildlife, bicycles, and human drivers. It fails to recognize potholes, detours, and roads with worn markings. Moments when harming a pedestrian could put its passenger at risk confuse driverless cars, and driverless cars have been found to have a higher crash rate than conventional vehicles. Driverless cars are also highly susceptible to hacking, whether through high-tech adversarial machine learning or low-tech graffiti.
Perhaps most frightening among these dangers is how driverless cars will interact with the dangers already present in our cities, like the fact that drivers are less likely to yield to people of color. With artificial intelligence already proven to amplify our biases, and face recognition software that struggles with darker skin, will driverless cars also be less likely to recognize a person of color waiting to cross the street?
Even if all these kinks get worked out, which seems unlikely, the benefits often promised to accompany the introduction of autonomous technology are far from guaranteed. The smartest driverless car will still consume more street space, risk more lives, and cause more harm per passenger trip than a bus ride, a bike ride, or a walk. The ascendency of for-hire vehicles and the decline of bus ridership has shown us that the majority of tomorrow’s driverless car passengers are today’s public transit riders — and since public transit is already 95% safer than driving a car, autonomous vehicles will have to be practically perfect to provide any holistic gains.
Will the dangers associated with autonomous vehicles keep them off the streets — or, as with conventional automobiles, will the blame be pushed off onto victims? Already, signs of a consequential future are trickling in: technologists have adopted the phrase “pedestrian interference,” as though walking is a glitch in the system, and the portmanteau “pedextrian” is in rising use, implying that the more than 10,000 pedestrians struck by drivers in New York City every year are the ones at fault. It is no coincidence that this new language of dangerous pedestrians is being deployed in the era of driverless cars, because the alternative — holding driverless cars to a higher safety standard than conventional ones — might not be in the best interests of the industry. As Sarah Kaufman at the New York University Rudin Center for Transportation warned, “Once our pedestrians realize these cars are programmed to stop when they cross the streets, there will be a jaywalking paradise and these cars will never get anywhere.”
A jaywalking paradise! Now that sounds like a future that TransAlt can get behind.
When cars first barrelled into cities a hundred years ago, our street were still playgrounds and marketplaces — a jaywalking paradise, you could say, except the term “jaywalker” had yet to be invented by automakers. It was a moment when companies like General Motors spent large amounts of capital convincing people that a car was a necessary and helpful urban technology, and that crossing the street without the permission of another new technology, the traffic light, was foolish and illegal. In the years that followed, in exchange for that technology, New York was introduced to widespread asthma and obesity, the death and injury of hundreds of thousands of its citizens in traffic crashes, and the loss of the vast majority of the city’s common, collective space.
The debate around autonomous vehicles is often framed as a question of whether New York City’s future will continue to be driven by these conventional automobiles or will be overtaken by driverless cars. This is a false choice. It disregards the fact that the vast majority of New Yorkers choose buses and subways for transportation, that more and more New Yorkers are taking to cycling, and that every New Yorker is a pedestrian at some point in their day. The question is not whether New Yorkers will choose autonomous or conventional vehicles, but whether our cities should have cars at all.
The question is not whether New Yorkers will choose autonomous or conventional vehicles, but whether our cities should have cars at all.
Driverless cars have been a hot topic since the late 1950s, when Arthur Radebaugh drew Closer Than You Think, the syndicated Sunday comic strip seen on these pages — but never more so than today, when the technology is revving toward reality. That conversation, however, is largely led by the people who stand to profit from autonomous technology, and for that reason, it asks all the wrong questions. Despite think pieces on the Trolley Problem and op-eds extolling the safety of driverless cars as an answer to our dangerous roads, we are failing to discuss the most important question: whether we want cars — autonomous or not — in our cities at all.
It won’t be long before General Motors decides that enough water has passed under the Brooklyn Bridge to move their autonomous vehicle testing schedule forward. Do you want to live in a driverless city full of cars, or in a city reclaimed from the automobile? Right now, for a little longer, it’s up to us.